软测量技术的研究及其在电石生产中的应用
|
摘要
在电石生产过程中,炉料电阻值的控制是稳定电石质量的关键。炉料电阻主要由炉内电石的比电阻决定,要控制炉料的电阻值,需要对电石的比电阻进行在线测量或估计。然而,至今国内外尚未开发出准确、可靠、廉价的冶炼炉在线测量仪表。因此,研究电石生产中比电阻值的软测量技术具有很大的理论意义与实用价值。
首先,本文在对矿热炉控制系统分析的基础上,给出了电石炉等效电路图,从几何学的角度推导了电石炉有功功率取最大值时炉料电阻的大小。其次,对软测量中数据的预处理进行了深入的研究与讨论,提出了数据的预处理应当分为实时预处理与离线预处理,并给出详细应用框图。然后,分析了最小二乘法的优缺点,并且详细介绍了如何利用渐消记忆的递推算法建立比电阻模型。接着,在分析实际工业过程中工况变化而导致模型精度下降的基础上,结合系数修正法与相关性的原理提出了一种新的在线校正方法。最后,分析模型失效的原因,根据辅助变量与主导变量的相关性强弱,给出了两种离线校正规则。
总而言之,本论文以电石生产中比电阻的软测量为基础,在测量数据预处理、最小二乘法、模型的建立及模型校正等方面进行了深入的研究,取得了有益的成果。这些成果对软测量技术,特别是电石生产中比电阻的软测量技术,起到了促进、丰富与深化的作用。同时,本论文的方法及结论,对于其它工业对象的软测量技术研究也具有一定的参考价值。
In the industrial process of calcium carbide manufacture,the resistance of charging stock control is the key to stabilize the quality of calcium carbide.The resistance of charging stock is determined by the resistivity of the calcium carbide.In order to control the value of the resistance of charging stock,the resistivity must be measured or estimated online.However, up to the present,the online measurement instrument of smelted furnace,which is precision, dependable,low cost and commercial,has not been developed throughout at home and abroad. Therefore,it is significant in theory and application to develop the soft sensing technology of the resistivity value in the industrial process of calcium carbide manufacture.
In the first place,based on analysis the submerged arc furnace control system,the circuit diagram of calcium carbide furnace is presented and the value of resistance of charging stock under the optimization power is proved by geometry method.In the second place,the data preparation should be made up of online preparation and offline preparation has been presented.In the same time,the application flow chart has been given.Then,the least squares especially the recursive least square based on lethe gene is studied and the model of resistivity is built.In the follow place,based on analysis the precision descended of the model in the practice industrial process,a new method for online correction is presented which integrated the correction factors and the correlation coefficients.At the last place,based on the reasons of the model invalidation and the value of correlation coefficients between secondary variable and the primary variable,two offline correction rules is presented.
In conclusion,this dissertation applied its main content to research and application of soft sensing technology of the resistivity value in the industrial process of calcium carbide manufacture.In some respects,such as data preprocessing,the least squares,the soft sensing model build,model correction and so on,has been lucubrated and some creative and helpful results are achieved after research work.The soft sensing technology,especially for the calcium carbide manufacture,should be enhanced,enriched and deepened by these results.In the same,these results should have reterence to soft sensor in other industrial process.
首先,本文在对矿热炉控制系统分析的基础上,给出了电石炉等效电路图,从几何学的角度推导了电石炉有功功率取最大值时炉料电阻的大小。其次,对软测量中数据的预处理进行了深入的研究与讨论,提出了数据的预处理应当分为实时预处理与离线预处理,并给出详细应用框图。然后,分析了最小二乘法的优缺点,并且详细介绍了如何利用渐消记忆的递推算法建立比电阻模型。接着,在分析实际工业过程中工况变化而导致模型精度下降的基础上,结合系数修正法与相关性的原理提出了一种新的在线校正方法。最后,分析模型失效的原因,根据辅助变量与主导变量的相关性强弱,给出了两种离线校正规则。
总而言之,本论文以电石生产中比电阻的软测量为基础,在测量数据预处理、最小二乘法、模型的建立及模型校正等方面进行了深入的研究,取得了有益的成果。这些成果对软测量技术,特别是电石生产中比电阻的软测量技术,起到了促进、丰富与深化的作用。同时,本论文的方法及结论,对于其它工业对象的软测量技术研究也具有一定的参考价值。
In the industrial process of calcium carbide manufacture,the resistance of charging stock control is the key to stabilize the quality of calcium carbide.The resistance of charging stock is determined by the resistivity of the calcium carbide.In order to control the value of the resistance of charging stock,the resistivity must be measured or estimated online.However, up to the present,the online measurement instrument of smelted furnace,which is precision, dependable,low cost and commercial,has not been developed throughout at home and abroad. Therefore,it is significant in theory and application to develop the soft sensing technology of the resistivity value in the industrial process of calcium carbide manufacture.
In the first place,based on analysis the submerged arc furnace control system,the circuit diagram of calcium carbide furnace is presented and the value of resistance of charging stock under the optimization power is proved by geometry method.In the second place,the data preparation should be made up of online preparation and offline preparation has been presented.In the same time,the application flow chart has been given.Then,the least squares especially the recursive least square based on lethe gene is studied and the model of resistivity is built.In the follow place,based on analysis the precision descended of the model in the practice industrial process,a new method for online correction is presented which integrated the correction factors and the correlation coefficients.At the last place,based on the reasons of the model invalidation and the value of correlation coefficients between secondary variable and the primary variable,two offline correction rules is presented.
In conclusion,this dissertation applied its main content to research and application of soft sensing technology of the resistivity value in the industrial process of calcium carbide manufacture.In some respects,such as data preprocessing,the least squares,the soft sensing model build,model correction and so on,has been lucubrated and some creative and helpful results are achieved after research work.The soft sensing technology,especially for the calcium carbide manufacture,should be enhanced,enriched and deepened by these results.In the same,these results should have reterence to soft sensor in other industrial process.
引文
[1]王炳盛,刘亚奎.电石生产的发展方向[J].中国氯碱,2007,5:43-44
[2]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,2000:1
[3]C.B.Brosillow.Inferential Control of Process[J].Journal of American Institute of Chemical Engineers.1978,24(3):485-509
[4]Dong Dong,Thomas J.McAvoy.Emission Monitoring Using Multivariate Soft Sensors[J].Proceedings of the American Control Conference Seattle.1995,35(2):761-765
[5]D.Wang,R.Srinivasan,J.Liu,P.N.S.Guru,K.M.Leong.Data-driven Soft Sensor Approach For Quality Prediction in a Refinery Process[J].IEEE,International Conference on Industrial Informatics,2006,230-235
[6]化工百科全书[M].第3卷.北京:化学工业出版,1993:439-446
[7]蔡杰.我国电石行业现状及发展建议[J].化工技术经济,2006,24(9):6-13
[8]熊谟远.电石生产及其深加工产品[M].北京:化学工业出版社,1989:8-9
[9]Mcacvoy T J.Contemplative Stance for Chemical Process Control[J].Automatica,1992,28(2):441-442
[10]李海青,黄志尧.软测量技术原理及应用[M].北京:化学工业出版社,2000:2-28
[11]Lee.J,Morari.M.Robust Measurement Selection[J].Automatica.1991,25(3):519-527
[12]罗荣富,邵惠鹤.推断控制中二次变量选择方法的研究[J].1992年控制与决策学术年会论文集[C].控制与决策编委会,1992
[13]罗荣富,邵惠鹤.软测量方法及其工业应用[J].第六届过程控制科学报告会论文集[C].上海:上海交通大学出版社,1993:324-329
[14]Narasimhan S,Mah R.S.H.Generalized Likelihood Ration Method or Gross Eroors Identification[J].Journal of American Institute of Chemical Engineers.1987,33(9):1514-1521
[15]Hongwei Tong,Cameron M Crorve.Detection Persistent Gross Errors by Sequential of Principal Components[J].Journal of American Institute of Chemical Engineers.1997,43(5):1242-1249
[16]Alex Kalosa,Arthur Kordon,Guido Smitsb,Sofka Werkmeister.Hybrid Model Development Methodology for Industrial Soft Sensors.Proceedings of the American Control Conference.2003.6:5417-5422
[17]G.D.Gonz a lez.Soft sensors for processing plants.Intelligent Processing and Manufacturing of Materials,1999,7:59-69
[18]L.Fortuna,S.Graziani,M.G.Xibilia,G.Napoli.Comparing Regressors Selection Methods for the Soft Sensor Design of a Sulfur Recovery Unit.Control and Automation,2006,6:1-6
[19]王金林,邵之江,张仲广,钱积新.乙醛生产过程中的软测量实现[J].化工自动化及仪表,2004,31(1):56-58
[20]李永博,孙瑜.电石生产中比电阻的软测量[J].化工自动化及仪表,2005,32(1):21-22
[21]C.M.Bo,J.Li,S.Zhang,C.Y.Sun,Y.R.Wang.The Application of Neural Network Soft Sensor Technology to an Advanced Control System of Distillation Operation.Neural Networks,2003,7(2):1054-1058
[22]Dirk Devogelaere,Marcel Rijckaert,Osvaldo Goza Leon,Gil Cruz Lemus.Application of Feedforward Neural Networks for Soft Sensors in the Sugar Industry.Neural Networks,2002,11:2-6
[23]罗晓,陈耀,孙优贤.基于统计回归的质量推断方法[J].信息与控制,2001,30(5):422-426
[24]颜学峰,余娟,钱锋.基于自适应偏最小二乘回归的初顶石脑油干点软测量[J].化工学报,2005,58(8):1511-1515
[25]翟军勇,费树岷,张湜.软测量技术和约束控制在精馏塔优化控制中的应用[J].工业仪表与自动化装置,2004,2:24-26
[26]上海吴淞化工化工厂《电石生产》业余写作组.电石生产[M].北京:燃料化学工业出版社,1972:4-5
[27]Rettkowski,W.,Geihufe,Chr.,Rudiger,K.H.,Chemical Technology,1976,28:588
[28]JI.A.库兹涅佐夫.电石生产[M].北京:化学工业出版社,1958:8-16
[29]府谷县新龙化工有限责任公司编印.12000KVA电石炉生产规程[M].2001
[30]杨忠魁.矿热炉电路分析计算[J].冶金动力,1995,4:16-24
[31]孙优贤,褚健.工业过程控制技术方法篇[M].北京:化学工业出版社,2006:354-376
[32]朱军.线性模型分析原理[M].北京:科学出版社,1999:48-49
[33]梁林.基于非线性部分最小二乘的软测量建模方法研究[D].北京:清华大学,2000,12
[34]马明建.数据采集与处理技术[M].西安:西安交通大学出版社,2005:215-231
[35]肖明耀.实验误差估计与数据处理[M].北京:科学出版社,1980:142-144
[36]王正光,周忠英,侯伯亨,李伯成.数据采集与处理[M].北京:国防工业出版社, 1985:160-181
[37]李艳.制浆蒸煮过程纸浆卡伯值软测量技术研究与应用[D].广东:华南理工大学,2003.5
[38]http://www.bjx.com.cn/files/wx/gyybyzdhzz/2005-5/2.htm
[39]马立平.统计数据标准化——无量纲化方法[J].北京统计,2000,3:34-35
[40]陈望春,赵立锋.权函数法在P-Ⅲ型分布中的应用[J].浙江水利科技,2002,4:46-47
[41]陈铁山.软测量技术在造纸水分控制中的应用研究[D].南京林业大学,2005
[42]王秀峰,卢桂章.系统建模与辨识[M].北京:电子工业出版社,2004:29-31
[43]刘钦圣.最小二乘问题计算方法[M].北京:北京工业大学出版社,1989:21-25
[44]Ji Wang,Rongxing Wu,Wenhua Zhao,Jianke Du,Dejin Huang.Correction Factors for Mindlin Higher-order Plate Theory with the Consideration of Electrodes[J].Frequency Control Symposium,2007,5:203-207
[45]Moulthrop,A.A.,Muha M.S..Accurate measurement of signals close to the noise floor on a spectrum analyzer[J].Microwave Theory and Techniques,1991,39(11):1882-1885
[46]Field H.,Emery K..An uncertainty analysis of the spectral correction factor[J].Photovoltaic Specialists Conference,1993,5:1180-1187
[47]Roshen W.Iron losses in permanent magnet synchronous motors[J].Industrial Electronics Society,2005,11:4
[48]William V.Nicholson.Object detection by correlation coefficients using azimuthally averaged reference projections[J].Biomedical Engineering,51(11):2006-2012
[49]Shuangquan Wang.Envelope Correlation Coefficient for Logarithmic Diversity Receivers Revisited[J].Communications,2007,11:2042-2046
[50]王村青.先进控制技术及应用[M].北京:化学工业出版社,2001:42
[51]杨善清,金涛.电石生产过程的优化控制[J].节能,1998,11:18-20
[2]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,2000:1
[3]C.B.Brosillow.Inferential Control of Process[J].Journal of American Institute of Chemical Engineers.1978,24(3):485-509
[4]Dong Dong,Thomas J.McAvoy.Emission Monitoring Using Multivariate Soft Sensors[J].Proceedings of the American Control Conference Seattle.1995,35(2):761-765
[5]D.Wang,R.Srinivasan,J.Liu,P.N.S.Guru,K.M.Leong.Data-driven Soft Sensor Approach For Quality Prediction in a Refinery Process[J].IEEE,International Conference on Industrial Informatics,2006,230-235
[6]化工百科全书[M].第3卷.北京:化学工业出版,1993:439-446
[7]蔡杰.我国电石行业现状及发展建议[J].化工技术经济,2006,24(9):6-13
[8]熊谟远.电石生产及其深加工产品[M].北京:化学工业出版社,1989:8-9
[9]Mcacvoy T J.Contemplative Stance for Chemical Process Control[J].Automatica,1992,28(2):441-442
[10]李海青,黄志尧.软测量技术原理及应用[M].北京:化学工业出版社,2000:2-28
[11]Lee.J,Morari.M.Robust Measurement Selection[J].Automatica.1991,25(3):519-527
[12]罗荣富,邵惠鹤.推断控制中二次变量选择方法的研究[J].1992年控制与决策学术年会论文集[C].控制与决策编委会,1992
[13]罗荣富,邵惠鹤.软测量方法及其工业应用[J].第六届过程控制科学报告会论文集[C].上海:上海交通大学出版社,1993:324-329
[14]Narasimhan S,Mah R.S.H.Generalized Likelihood Ration Method or Gross Eroors Identification[J].Journal of American Institute of Chemical Engineers.1987,33(9):1514-1521
[15]Hongwei Tong,Cameron M Crorve.Detection Persistent Gross Errors by Sequential of Principal Components[J].Journal of American Institute of Chemical Engineers.1997,43(5):1242-1249
[16]Alex Kalosa,Arthur Kordon,Guido Smitsb,Sofka Werkmeister.Hybrid Model Development Methodology for Industrial Soft Sensors.Proceedings of the American Control Conference.2003.6:5417-5422
[17]G.D.Gonz a lez.Soft sensors for processing plants.Intelligent Processing and Manufacturing of Materials,1999,7:59-69
[18]L.Fortuna,S.Graziani,M.G.Xibilia,G.Napoli.Comparing Regressors Selection Methods for the Soft Sensor Design of a Sulfur Recovery Unit.Control and Automation,2006,6:1-6
[19]王金林,邵之江,张仲广,钱积新.乙醛生产过程中的软测量实现[J].化工自动化及仪表,2004,31(1):56-58
[20]李永博,孙瑜.电石生产中比电阻的软测量[J].化工自动化及仪表,2005,32(1):21-22
[21]C.M.Bo,J.Li,S.Zhang,C.Y.Sun,Y.R.Wang.The Application of Neural Network Soft Sensor Technology to an Advanced Control System of Distillation Operation.Neural Networks,2003,7(2):1054-1058
[22]Dirk Devogelaere,Marcel Rijckaert,Osvaldo Goza Leon,Gil Cruz Lemus.Application of Feedforward Neural Networks for Soft Sensors in the Sugar Industry.Neural Networks,2002,11:2-6
[23]罗晓,陈耀,孙优贤.基于统计回归的质量推断方法[J].信息与控制,2001,30(5):422-426
[24]颜学峰,余娟,钱锋.基于自适应偏最小二乘回归的初顶石脑油干点软测量[J].化工学报,2005,58(8):1511-1515
[25]翟军勇,费树岷,张湜.软测量技术和约束控制在精馏塔优化控制中的应用[J].工业仪表与自动化装置,2004,2:24-26
[26]上海吴淞化工化工厂《电石生产》业余写作组.电石生产[M].北京:燃料化学工业出版社,1972:4-5
[27]Rettkowski,W.,Geihufe,Chr.,Rudiger,K.H.,Chemical Technology,1976,28:588
[28]JI.A.库兹涅佐夫.电石生产[M].北京:化学工业出版社,1958:8-16
[29]府谷县新龙化工有限责任公司编印.12000KVA电石炉生产规程[M].2001
[30]杨忠魁.矿热炉电路分析计算[J].冶金动力,1995,4:16-24
[31]孙优贤,褚健.工业过程控制技术方法篇[M].北京:化学工业出版社,2006:354-376
[32]朱军.线性模型分析原理[M].北京:科学出版社,1999:48-49
[33]梁林.基于非线性部分最小二乘的软测量建模方法研究[D].北京:清华大学,2000,12
[34]马明建.数据采集与处理技术[M].西安:西安交通大学出版社,2005:215-231
[35]肖明耀.实验误差估计与数据处理[M].北京:科学出版社,1980:142-144
[36]王正光,周忠英,侯伯亨,李伯成.数据采集与处理[M].北京:国防工业出版社, 1985:160-181
[37]李艳.制浆蒸煮过程纸浆卡伯值软测量技术研究与应用[D].广东:华南理工大学,2003.5
[38]http://www.bjx.com.cn/files/wx/gyybyzdhzz/2005-5/2.htm
[39]马立平.统计数据标准化——无量纲化方法[J].北京统计,2000,3:34-35
[40]陈望春,赵立锋.权函数法在P-Ⅲ型分布中的应用[J].浙江水利科技,2002,4:46-47
[41]陈铁山.软测量技术在造纸水分控制中的应用研究[D].南京林业大学,2005
[42]王秀峰,卢桂章.系统建模与辨识[M].北京:电子工业出版社,2004:29-31
[43]刘钦圣.最小二乘问题计算方法[M].北京:北京工业大学出版社,1989:21-25
[44]Ji Wang,Rongxing Wu,Wenhua Zhao,Jianke Du,Dejin Huang.Correction Factors for Mindlin Higher-order Plate Theory with the Consideration of Electrodes[J].Frequency Control Symposium,2007,5:203-207
[45]Moulthrop,A.A.,Muha M.S..Accurate measurement of signals close to the noise floor on a spectrum analyzer[J].Microwave Theory and Techniques,1991,39(11):1882-1885
[46]Field H.,Emery K..An uncertainty analysis of the spectral correction factor[J].Photovoltaic Specialists Conference,1993,5:1180-1187
[47]Roshen W.Iron losses in permanent magnet synchronous motors[J].Industrial Electronics Society,2005,11:4
[48]William V.Nicholson.Object detection by correlation coefficients using azimuthally averaged reference projections[J].Biomedical Engineering,51(11):2006-2012
[49]Shuangquan Wang.Envelope Correlation Coefficient for Logarithmic Diversity Receivers Revisited[J].Communications,2007,11:2042-2046
[50]王村青.先进控制技术及应用[M].北京:化学工业出版社,2001:42
[51]杨善清,金涛.电石生产过程的优化控制[J].节能,1998,11:18-20