基于Agent的主分馏塔粗汽油产品质量控制及仿真
|
摘要
反映流化催化裂化装置(FCCU)主分馏塔产品质量指标的重要参数是粗汽油的干点和轻柴油的倾点。对它们实行“卡边控制”,提高经济效益,一直都是石油炼制企业所关注的问题。在加工石油过程中,催化裂化主分馏塔系统的机理比较复杂、耦合严重、非线性强以及干扰因素众多,以致很难建立适当的模型,并且装置的产品质量、收率等指标,往往受到严格的操作条件制约。所以,采用传统的控制策略很难平稳运行装置。针对主分馏塔系统,寻找有效增产、节能的先进控制和过程优化方法是过程控制领域中富有挑战性的工作。
本文以提高产品质量为目标,运用基于Agent网络模型的控制方法对塔顶分馏过程进行控制,以粗汽油产品质量控制为切入点,展开了理论与实践两方面的研究,主要内容如下:
1.针对复杂系统的过程控制问题,提出了一种基于智能Agent的Agent网络模型方法。该方法通过预估和控制算法对复杂工业过程进行预估和控制,增强了系统的稳定性。
2.认真分析催化裂化主分馏塔的工艺流程,找到影响粗汽油产品质量的主要因素以及它们之间的控制关系和控制难点。利用质量变量与加工过程机理间的关系建立测量模型,通过建立的粗汽油干点软测量分析仪,实施产品质量的监测与控制。
3.根据影响粗汽油产品质量的主要因素以及它们之间的控制关系,按照Agent网络模型思想,提取控制方面起主要作用的控制关系,建立粗汽油产品质量控制的Agent网络模型。然后,对该模型的预估和控制算法进行设计。系统分析和设计,均在各Agent单元级进行。整体系统控制品质的获得依赖于各Agent单元按照结构的协调,控制指令从上级逐级向下级传递,这样保证了各Agent单元的设计与子网络和整体性能联系起来。
仿真结果显示,本文建立的粗汽油产品质量控制的Agent网络模型,具有良好的稳定性、抗干扰和容错能力,实现了粗汽油产品质量的有效控制。
Crude gasoline endpoint and light diesel oil pour point is important parameters to reflect product quality indicators of FCCU main fractionators. Its implementation of "bounder control" to enhance economic efficiency has been the oil refining enterprises concern. In the oil processing, mechanism of the FCCU main fractionating tower is complex; nonlinearity is strong; coupling is serious, and the number of interference factor is numerous. So it is difficult to establish a satisfactory model. Moreover, targets including yield coefficient of the device and product quality are often restricted by the harsh operating condition. It is very difficult to realize smooth operation of the device by conventional control strategy. For the main fractionating tower system, looking for effective increase production, energy-saving advanced control and process optimization method is a challenging job in the field of process control.
This paper takes improving the quality of products as the objective, uses control method based Agent network model to control the fractionation process in the top tower, takes crude gasoline product quality control as the breakthrough point, and carries out both the theory and practice of research.
The main contents are as follows:
1. For process control problem of the complex system, a Multi-Agent-based network model of the Agent was put forward. This method was used to prospect and control the complex industrial process through prediction and control algorithm, enhanced the stability of the system.
2. Carefully analyzed process flow of the FCCU main fractionators, succeeded in finding the major factors, impacted crude gasoline product quality, as well as their control relationship and control difficult points. Using the relationship of quality variables and process variables, or the relationship between quality variables and processing mechanisms, set up the measurement model. Then monitoring and control the product quality through the established crude gasoline endpoint soft-sensing analyzer.
3. According to the major factors, impact crude gasoline product quality, as well as the control relationship between them, in accordance with the Agent network model idea, extract the control relationship, play a major role in control system, establish the Agent network model of crude gasoline product for quality control. Then design prediction and control algorithm of the model. Systems analysis and design, all go ahead in the Agent units. The control quality of the overall system was dependent on the Agent units in accordance with the structure of coordination, control commands from the superior to the lower level transmission; this ensures that the design of the Agent units can be linked with control of sub-networks and the overall performance.
The simulation results show that the Agent-based network model makes the qualitycontrol of crude gasoline endpoint has a good stability, tolerance and anti-interferencecapability, realized the effective control of the product quality.
本文以提高产品质量为目标,运用基于Agent网络模型的控制方法对塔顶分馏过程进行控制,以粗汽油产品质量控制为切入点,展开了理论与实践两方面的研究,主要内容如下:
1.针对复杂系统的过程控制问题,提出了一种基于智能Agent的Agent网络模型方法。该方法通过预估和控制算法对复杂工业过程进行预估和控制,增强了系统的稳定性。
2.认真分析催化裂化主分馏塔的工艺流程,找到影响粗汽油产品质量的主要因素以及它们之间的控制关系和控制难点。利用质量变量与加工过程机理间的关系建立测量模型,通过建立的粗汽油干点软测量分析仪,实施产品质量的监测与控制。
3.根据影响粗汽油产品质量的主要因素以及它们之间的控制关系,按照Agent网络模型思想,提取控制方面起主要作用的控制关系,建立粗汽油产品质量控制的Agent网络模型。然后,对该模型的预估和控制算法进行设计。系统分析和设计,均在各Agent单元级进行。整体系统控制品质的获得依赖于各Agent单元按照结构的协调,控制指令从上级逐级向下级传递,这样保证了各Agent单元的设计与子网络和整体性能联系起来。
仿真结果显示,本文建立的粗汽油产品质量控制的Agent网络模型,具有良好的稳定性、抗干扰和容错能力,实现了粗汽油产品质量的有效控制。
Crude gasoline endpoint and light diesel oil pour point is important parameters to reflect product quality indicators of FCCU main fractionators. Its implementation of "bounder control" to enhance economic efficiency has been the oil refining enterprises concern. In the oil processing, mechanism of the FCCU main fractionating tower is complex; nonlinearity is strong; coupling is serious, and the number of interference factor is numerous. So it is difficult to establish a satisfactory model. Moreover, targets including yield coefficient of the device and product quality are often restricted by the harsh operating condition. It is very difficult to realize smooth operation of the device by conventional control strategy. For the main fractionating tower system, looking for effective increase production, energy-saving advanced control and process optimization method is a challenging job in the field of process control.
This paper takes improving the quality of products as the objective, uses control method based Agent network model to control the fractionation process in the top tower, takes crude gasoline product quality control as the breakthrough point, and carries out both the theory and practice of research.
The main contents are as follows:
1. For process control problem of the complex system, a Multi-Agent-based network model of the Agent was put forward. This method was used to prospect and control the complex industrial process through prediction and control algorithm, enhanced the stability of the system.
2. Carefully analyzed process flow of the FCCU main fractionators, succeeded in finding the major factors, impacted crude gasoline product quality, as well as their control relationship and control difficult points. Using the relationship of quality variables and process variables, or the relationship between quality variables and processing mechanisms, set up the measurement model. Then monitoring and control the product quality through the established crude gasoline endpoint soft-sensing analyzer.
3. According to the major factors, impact crude gasoline product quality, as well as the control relationship between them, in accordance with the Agent network model idea, extract the control relationship, play a major role in control system, establish the Agent network model of crude gasoline product for quality control. Then design prediction and control algorithm of the model. Systems analysis and design, all go ahead in the Agent units. The control quality of the overall system was dependent on the Agent units in accordance with the structure of coordination, control commands from the superior to the lower level transmission; this ensures that the design of the Agent units can be linked with control of sub-networks and the overall performance.
The simulation results show that the Agent-based network model makes the qualitycontrol of crude gasoline endpoint has a good stability, tolerance and anti-interferencecapability, realized the effective control of the product quality.
引文
[1]工业过程先进控制Industrial Advanced Process Control来源:SUPCON企业综合自动化论坛
[2]解怀仁,杨彬彦.石油化工仪表控制系统选用手册[M].北京:中国石化出版社,2004
[3]Liu Qizhong and Kang Qingshan.The implementation of FCCU main fractionator tray 20 temperance feed forward control system[J].Automation in Refined and Chemical Industry.1997,(5):29-31
[4]华宏坚,张遥,陈立.充分利用DCS功能提高催化裂化装置运行质量.中国工控网
[5]王树青,金晓明等著.先进控制技术应用实例[M].北京:化学工业出版社,2005
[6]鲍英军,王景成,竺建敏.DMCplus控制器在催化裂化装置的应用[J].炼油设计与工程.2003,33(6):24-27
[7]洛阳石化公司,石油大学.“催化裂化装置反再分馏系统先进控制”鉴定材料,1996
[8]张克进,沈雁鸣,俞金寿.软测量在分馏塔粗汽油干点推断控制中的应用[J].自动化仪表,1999,20(11):17-19
[9]钟璇,张泉灵,王树青.催化裂化主分馏塔产品质量的广义预测控制策略[J].控制理论与应用,2001,(18):134-140
[10]胡品慧,马庆春等.汽油产品质量先进控制与工程应用[J].化工自动化及仪表.2003,30(4):23-25
[11]杨淑莹,郭翠梨.FCCU分馏塔产品质量预测系统的设计[J].哈尔滨工业大学学报.2005,37(4):501-503
[12]Holland J H.Hidden Order:How Adaptation Builds Complexity[J].Reading,MA:Addison-Wesley,1995
[13]Holland J H.Adaptation in Natural and Artificial Systems:An Introductory Analysis with Application to Biology[J],Control and Artificial Intelligence,2~(nd)ed.Cambridge,MA:MIT Press,1992
[14]Holland J H,Holyoak K J,Nisbett R E,Thagard P R.Processes of Inference, Learning,and Discovery[J].Cambridge,MA:MIT Press,1986
[15]Murata T.Petri-nets:properties,analysis and application[A].Proc IEEE 77[C].1989:541-579
[16]刘金琨,王树青.复杂工业过程多智能体协调Petri网模型[J].系统仿真学报.1999,11(5):361-363
[17]陈铁军.链系统方法及其运用[M].郑州:河南科学技术出版社,1993
[18]陈铁军,邱祖廉.链控制器及其应用[J].自动化学报,1994,20(3):379-382
[19]陈铁军.结构分散化控制理论[M].新疆:科技出版社.2004
[20]JAN WILLEMS,The Behavioral Approach to Open and Interconnected Systems [J],IEEE CONTROL SYSTEMS MAGAZINE(S0272-1708),2007,(12):46-99
[21]廖守亿,戴金海.复杂适应系统及基于Agent的建模与仿真方法[J].系统仿真学报.2004,16(1):113-117
[22]王汝传,徐小龙,黄海平.智能Agent及其在信息网络中的应用[M].北京:北京邮电大学出版社,2006
[23]廖守亿.复杂系统基于Agent的建模与仿真方法研究及应用[D].国防科学技术大学硕士学位论文.2005
[24]W.van der Hoek and M.Wboldridge.Model Checking Cooperation,Knowledge,and Time-A Case Study[J],In Research In Economics,2003,57(3):235-265
[25]Y.Ono,H.Uchiyama and W.Potter,A mobile robot for corridor navigation:A multi-agent approach[C],ACMSE'04:ACM Southeast Regional Conference,ACM Press(2004):379-384
[26]Rieser,M.and Nagel,K.Network breakdown “at the edge of chaos” in multi-agent traffic simulations[J].European Physical Journal B,2008,63(3):321-327
[27]Madureira,Ana and Santos,Joaquim.Proposal of multi-agent based model for dynamic scheduling in manufacturing[J].WSEAS Transactions on Information Science and Applications.2005,2(5):600-605
[28]Lin,Jim-Min;Yang,Hongji and Fang,Guo-Ming.etc.A multi-agent distributed scripting system for COTS-based distributed software integration[J].Tamkang Journal of Science and Engineering,2007,10(4):351-362
[29]窦立谦.多Agent系统协作与协调的研究[D].天津大学硕士学位论文.2005
[30]陈玲,丁纪凯.基于Multi-Agent理论的复杂过程控制系统[J].机电一体化. 2004,(1):28-31
[31]黄琳,段志生.控制科学中的复杂性[J].自动化学报.2003,29(5):748-754
[32]郭一楠,巩敦卫,程健.基于多智能体的回炉煤气主管压力控制系统[J].信息与控制.2006,35(5):634-640
[33]李斌,师汉民,胡春华等.基于Agent分布式网络化制造模式的研究[J].中国机械工程.1999,10(12):1358-1362
[34]陈铁军,李新瑞.基于结构分散化控制的球磨机制粉系统控制与仿真[J].计算技术与自动化.2007,26(3):21-24
[35]徐国忠.催化裂化分馏塔粗汽油干点的测量与控制[J].石油化工自动化,2001,(3):31-33
[36]周文娟.催化裂化主分馏塔和吸收稳定系统工艺模拟与改进研究[D].天津大学硕士学位论文.2005
[37]CALDERON.Neural network based framework for fault diagnosis in batch chemical plants[J].Computers&Chemical Engineering,2000,24(2):777-781
[38]杨宝康,李庆萍,蔡永清.催化裂化装置培训教程[M].北京:化学工业出版社.2005
[39]林世雄.石油炼制工程[M].北京:石油工业出版社,1997
[40]梁凤印.流化催化裂化[M].北京:中国石化出版社.2006
[41]包健,曾文华,严义.基于神经网络的汽油干点软测量[J].杭州电子工业学院学报2000,20(1):72-75
[42]马伯文.催化裂化装置技术问答[M].中国石化出版社.2003
[43]郭锦标.MPC在重油催化裂化过程控制中的应用[J].炼油设计.2000,30(2):32-36
[44]李成栋.催化重整装置技术问答[M].北京:中国石化出版社.2004
[45]刘先广,黄德先,田学民.基于严格机理方法计算FCCU主分馏塔粗汽油干点[C].第十三届中国过程控制年会过程控制科学技术与应用(2002)
[46]刘齐忠,康庆山.FCCU分馏塔20层塔盘温度前馈控制系统在DCS上的实施[J].炼油化工自动化.1997,5:29-31
[47]赵森林.汽油干点预估模型与在线控制[J].石化技术与应用.2000,18(4):221-222
[48]Contreras,Miguel;Sheremetov,Leonid.Industrial application integration using the unification approach to agent-enabled semantic SOA.Robotics and Computer- Integrated Manufacturing[J],2008,24(5):680-695
[49]Lim,Sun-Jong;Song,Jun.Yeob.Application of the agent concept to e-beam manufacturing system[C].ICCAS 2007,International Conference on Control,Automation and Systems,2007:2186-2189
[50]Suganuma,Takuo;Sugawara,Kenji.Concept of symbiotic computing and its agent-based application to a ubiquitous care-support service[J].International Journal of Cognitive Informatics and Natural Intelligence,2009,3(1):34-56
[51]陈玲.基于Multi-Agent的复杂工业过程控制系统及其在PROFIBUS上实现的研究[D].东华大学硕士学位论文.2004
[52]Al-Kanhal,Tawfeeq;Abbod,Maysam.Multi-agent system for dynamic manufacturing system optimization[J].Lecture Notes in Computer Science.2008,3:634-643
[53]Fagui,Liu;Gong sheng,Zhang.PCE-oriented Multi-Agent software framework and its interactive computing application[C].Proceedings of 2008 3rd International Conference on Intelligent System and Knowledge Engineering,ISKE 2008,2008:170-175
[2]解怀仁,杨彬彦.石油化工仪表控制系统选用手册[M].北京:中国石化出版社,2004
[3]Liu Qizhong and Kang Qingshan.The implementation of FCCU main fractionator tray 20 temperance feed forward control system[J].Automation in Refined and Chemical Industry.1997,(5):29-31
[4]华宏坚,张遥,陈立.充分利用DCS功能提高催化裂化装置运行质量.中国工控网
[5]王树青,金晓明等著.先进控制技术应用实例[M].北京:化学工业出版社,2005
[6]鲍英军,王景成,竺建敏.DMCplus控制器在催化裂化装置的应用[J].炼油设计与工程.2003,33(6):24-27
[7]洛阳石化公司,石油大学.“催化裂化装置反再分馏系统先进控制”鉴定材料,1996
[8]张克进,沈雁鸣,俞金寿.软测量在分馏塔粗汽油干点推断控制中的应用[J].自动化仪表,1999,20(11):17-19
[9]钟璇,张泉灵,王树青.催化裂化主分馏塔产品质量的广义预测控制策略[J].控制理论与应用,2001,(18):134-140
[10]胡品慧,马庆春等.汽油产品质量先进控制与工程应用[J].化工自动化及仪表.2003,30(4):23-25
[11]杨淑莹,郭翠梨.FCCU分馏塔产品质量预测系统的设计[J].哈尔滨工业大学学报.2005,37(4):501-503
[12]Holland J H.Hidden Order:How Adaptation Builds Complexity[J].Reading,MA:Addison-Wesley,1995
[13]Holland J H.Adaptation in Natural and Artificial Systems:An Introductory Analysis with Application to Biology[J],Control and Artificial Intelligence,2~(nd)ed.Cambridge,MA:MIT Press,1992
[14]Holland J H,Holyoak K J,Nisbett R E,Thagard P R.Processes of Inference, Learning,and Discovery[J].Cambridge,MA:MIT Press,1986
[15]Murata T.Petri-nets:properties,analysis and application[A].Proc IEEE 77[C].1989:541-579
[16]刘金琨,王树青.复杂工业过程多智能体协调Petri网模型[J].系统仿真学报.1999,11(5):361-363
[17]陈铁军.链系统方法及其运用[M].郑州:河南科学技术出版社,1993
[18]陈铁军,邱祖廉.链控制器及其应用[J].自动化学报,1994,20(3):379-382
[19]陈铁军.结构分散化控制理论[M].新疆:科技出版社.2004
[20]JAN WILLEMS,The Behavioral Approach to Open and Interconnected Systems [J],IEEE CONTROL SYSTEMS MAGAZINE(S0272-1708),2007,(12):46-99
[21]廖守亿,戴金海.复杂适应系统及基于Agent的建模与仿真方法[J].系统仿真学报.2004,16(1):113-117
[22]王汝传,徐小龙,黄海平.智能Agent及其在信息网络中的应用[M].北京:北京邮电大学出版社,2006
[23]廖守亿.复杂系统基于Agent的建模与仿真方法研究及应用[D].国防科学技术大学硕士学位论文.2005
[24]W.van der Hoek and M.Wboldridge.Model Checking Cooperation,Knowledge,and Time-A Case Study[J],In Research In Economics,2003,57(3):235-265
[25]Y.Ono,H.Uchiyama and W.Potter,A mobile robot for corridor navigation:A multi-agent approach[C],ACMSE'04:ACM Southeast Regional Conference,ACM Press(2004):379-384
[26]Rieser,M.and Nagel,K.Network breakdown “at the edge of chaos” in multi-agent traffic simulations[J].European Physical Journal B,2008,63(3):321-327
[27]Madureira,Ana and Santos,Joaquim.Proposal of multi-agent based model for dynamic scheduling in manufacturing[J].WSEAS Transactions on Information Science and Applications.2005,2(5):600-605
[28]Lin,Jim-Min;Yang,Hongji and Fang,Guo-Ming.etc.A multi-agent distributed scripting system for COTS-based distributed software integration[J].Tamkang Journal of Science and Engineering,2007,10(4):351-362
[29]窦立谦.多Agent系统协作与协调的研究[D].天津大学硕士学位论文.2005
[30]陈玲,丁纪凯.基于Multi-Agent理论的复杂过程控制系统[J].机电一体化. 2004,(1):28-31
[31]黄琳,段志生.控制科学中的复杂性[J].自动化学报.2003,29(5):748-754
[32]郭一楠,巩敦卫,程健.基于多智能体的回炉煤气主管压力控制系统[J].信息与控制.2006,35(5):634-640
[33]李斌,师汉民,胡春华等.基于Agent分布式网络化制造模式的研究[J].中国机械工程.1999,10(12):1358-1362
[34]陈铁军,李新瑞.基于结构分散化控制的球磨机制粉系统控制与仿真[J].计算技术与自动化.2007,26(3):21-24
[35]徐国忠.催化裂化分馏塔粗汽油干点的测量与控制[J].石油化工自动化,2001,(3):31-33
[36]周文娟.催化裂化主分馏塔和吸收稳定系统工艺模拟与改进研究[D].天津大学硕士学位论文.2005
[37]CALDERON.Neural network based framework for fault diagnosis in batch chemical plants[J].Computers&Chemical Engineering,2000,24(2):777-781
[38]杨宝康,李庆萍,蔡永清.催化裂化装置培训教程[M].北京:化学工业出版社.2005
[39]林世雄.石油炼制工程[M].北京:石油工业出版社,1997
[40]梁凤印.流化催化裂化[M].北京:中国石化出版社.2006
[41]包健,曾文华,严义.基于神经网络的汽油干点软测量[J].杭州电子工业学院学报2000,20(1):72-75
[42]马伯文.催化裂化装置技术问答[M].中国石化出版社.2003
[43]郭锦标.MPC在重油催化裂化过程控制中的应用[J].炼油设计.2000,30(2):32-36
[44]李成栋.催化重整装置技术问答[M].北京:中国石化出版社.2004
[45]刘先广,黄德先,田学民.基于严格机理方法计算FCCU主分馏塔粗汽油干点[C].第十三届中国过程控制年会过程控制科学技术与应用(2002)
[46]刘齐忠,康庆山.FCCU分馏塔20层塔盘温度前馈控制系统在DCS上的实施[J].炼油化工自动化.1997,5:29-31
[47]赵森林.汽油干点预估模型与在线控制[J].石化技术与应用.2000,18(4):221-222
[48]Contreras,Miguel;Sheremetov,Leonid.Industrial application integration using the unification approach to agent-enabled semantic SOA.Robotics and Computer- Integrated Manufacturing[J],2008,24(5):680-695
[49]Lim,Sun-Jong;Song,Jun.Yeob.Application of the agent concept to e-beam manufacturing system[C].ICCAS 2007,International Conference on Control,Automation and Systems,2007:2186-2189
[50]Suganuma,Takuo;Sugawara,Kenji.Concept of symbiotic computing and its agent-based application to a ubiquitous care-support service[J].International Journal of Cognitive Informatics and Natural Intelligence,2009,3(1):34-56
[51]陈玲.基于Multi-Agent的复杂工业过程控制系统及其在PROFIBUS上实现的研究[D].东华大学硕士学位论文.2004
[52]Al-Kanhal,Tawfeeq;Abbod,Maysam.Multi-agent system for dynamic manufacturing system optimization[J].Lecture Notes in Computer Science.2008,3:634-643
[53]Fagui,Liu;Gong sheng,Zhang.PCE-oriented Multi-Agent software framework and its interactive computing application[C].Proceedings of 2008 3rd International Conference on Intelligent System and Knowledge Engineering,ISKE 2008,2008:170-175