基于遗传算法辨识的超临界机组给水控制
|
摘要
超临界锅炉是国家未来的发展方向,给水系统是其中的重要环节。超临界机组采用直流锅炉,从给水泵到汽机,汽水直接关联,使得锅炉各参数间和汽机与锅炉间具有强烈的耦合特性,加大了给水控制难度。在火电厂热工过程控制系统的分析设计中,普遍采用基于模型的控制方法,给水控制也不例外。本文以国电铜陵600MW超临界机组直流锅炉为例,通过对现场数据的分析和预处理,采用自适应遗传辨识方法建立100%及70%工况下的锅炉动态模型。针对多变量控制,提出了一种神经网络解耦控制方法,对建好的模型进行解耦,并用于中间点焓值修正煤水比的给水控制系统中。通过仿真可以看出该控制系统能有效克服扰动,保证中间点焓值,实现过热气温粗调。
Supercritical boilers will be the country's future development direction and water supply system is one of the important parts. Ultra-supercritical plants using once-through boiler, from the feed water pump to the steam machine, steam and water being directly related to each other, making each parameter between the boiler and steam turbine and boiler with strong coupling characteristics increase the difficulty of feedwater control. In the analysis and design of thermal power plant’s thermal process control system, model-based control method is widely used. This article taking the State Power Tongling 600MW supercritical once-through boiler for example, through on-site data analysis and pre-processing, dynamic model of the 100% and 70% boiler operating conditions is established using adaptive genetic identification method. According to multi-variable control, the article proposes a neural network decoupling control method that is used to decouple the built model, and is apllyed to the mid-point enthalpy correction coal-water ratio of the feedwater control system.Through the simulation, we can see that the control system can overcome disturbances to ensure the mid-point enthalpy value, to achieve an coarse regulation of the overheated temperature.
Supercritical boilers will be the country's future development direction and water supply system is one of the important parts. Ultra-supercritical plants using once-through boiler, from the feed water pump to the steam machine, steam and water being directly related to each other, making each parameter between the boiler and steam turbine and boiler with strong coupling characteristics increase the difficulty of feedwater control. In the analysis and design of thermal power plant’s thermal process control system, model-based control method is widely used. This article taking the State Power Tongling 600MW supercritical once-through boiler for example, through on-site data analysis and pre-processing, dynamic model of the 100% and 70% boiler operating conditions is established using adaptive genetic identification method. According to multi-variable control, the article proposes a neural network decoupling control method that is used to decouple the built model, and is apllyed to the mid-point enthalpy correction coal-water ratio of the feedwater control system.Through the simulation, we can see that the control system can overcome disturbances to ensure the mid-point enthalpy value, to achieve an coarse regulation of the overheated temperature.
引文
[1]樊泉桂.超超临界及亚临界参数锅炉.北京:中国电力出版社,2007.前言
[2]陈永阳,林荣文.浅析超临界机组给水系统.电站辅机,2005,4(95):16~18
[3]周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999
[4]雷德明.多维实数编码遗传算法.控制与决策,2000,15(3):239~241
[5]菅倩.基于矩阵编码遗传算法的研究与应用,[学位论文]:太原,太原理工大学,2008.5
[6] Schraudolph N N,Belew R K.Dynamic Parameter Encoding for Genetic Algorithms.Machine Learning,1992,9(1):9~21
[7]唐飞,滕弘飞,刘峻,等.带有能力约束的二维装填布局问题.宇航学报,2000,21(25)0~57
[8]张思才,张方晓.一种遗传算法适应度函数的改进方法.计算机应用与软件,2006,23(2):108~110
[9] S W Mahfoud.Genetic Drift in SharingMethods[C] .0-7803-1899-4/94,1994,IEEE,62~72
[10] V Petridis,SKazarlis.Varying Quality Function in Genetic Algorithms and the Cutting Problem[C].0-7803-1899-4/94,1994,IEEE,166~169
[11]杨旭东,张彤.遗传算法应用于系统在线辨识研究.哈尔滨工业大学学报,2000,32(1):102~105
[12] Back T,Hoffmeister F.Extended Election Mechanisms in Genetic Algorithms. In[49],1991:92~99
[13] Maza M D L,Tidor B.An Analysis of Selection Procedures with Particular Attention Paid to Proportion and Boltzmann Selection.In[50],1993:124~131
[14] D T Pham,G Jin.Genetic Algorithms Using Gradient-like reproduction operator. Electronics Letters,1995,31(18):1558~1559
[15] Michalewicz .ZGenetic Algorithms+Data structure=Evolution Programs[M] Springer-Verlag,Berlin1992
[16] Back T.Selection pressure in evolutionary algorithms:a Characterization of selection mechanisms[C].Proceeding of the First IEEE conference on Evolutionary computation,IEEE press Orlando,FL.1994:57~62
[17]邝航宇,金晶,苏勇.自适应遗传算法交叉变异算子的改进.计算机工程与应用,2006,12:93~96
[18]李良敏,温广瑞,王生昌.遗传算法中遗传操作的改进策略.计算机应用与软件, 2009, 26 (6):27~30
[19]张明辉,王尚锦.自适应搜索的改进遗传算法及其应用.西安交通大学学报,2002,36(3):226~229
[20] Feng Q X,Francesco P.Theoretical Analysis of Evolutionary Algorithms with an Infinite Population Size in Continuous Space Part I and PartⅡ:and Basic Properties of Selection and Mutation.IEEE Trans Neural Networks,1994,5(1):102~129
[21] Pan Z J,Kang L S,Nie S X.Evolving Both the Topology and Weights of Neural Network. Parallel Algorithms and Applications.1996,9:299~307
[22] Xie Nan,Chen Ying-jun.Improving strategies on genetic algorithm and its application to the optimum design of bridges under earthquake.Gong Cheng Li Xue /Engineering Mechanics 173,2000 Tsinghua Univ 31~36
[23] Wang Lei,Jiao Li-cheng.Novel genetic algorithm based on immunity[C]. Proceedings-IEEE International Symposium on Circuits and Systems 5 May 28-May 31 2000 Sponsored by:IEEE Circuits and Systems Society IEEE V385-V388
[24]骆晨钟,邵惠鹤.采用混沌变异的进化算法.控制与决策,2000,15(9):557~560
[25]叶在福,单渊达.基于多种群遗传算法的输电系统扩展规划.电力系统自动化,2000,10(5):24~27
[26]孟佳娜,王立宏.基于异构多种群的一种改进遗传算法.微电子学与计算机,2004,21(6):71~73
[27] Srinivas M,Patnaik L M.Adaptive probabilities of crossover and mutation in genetic algorithms,IEEE Trans on Systems,Man and Cybernetics,1994,24(4):656~667
[28]陈小平,于盛林.实数遗传算法交叉策略的改进.电子学报,2003,31(1):71~74
[29]王凌,黄璇,郑大钟.一类带筛选策略的改进遗传算法及其性能分析.控制与决策,2004,19(11):1290~1293
[30]杨金明,吴捷,钟丹虹.多目标优化问题中一种改进的遗传算法.华南理工大学学报(自然科学版),2001,29(12):65~67
[31]姚俊峰,梅炽,彭小奇.混沌遗传算法(CGA)的应用研究及其优化效率评价[J].自动化学报,2002,28(6):935~942
[32]王正志,薄涛.进化计算.国防科技大学出版社,2000
[33] Back T,Hoffmeister F.Extended Election Mechanismsin Genetic Algorithms. In[49],1991:92~99
[34] Maza M D L,Tidor B.An Analysis of Selection Procedures with Particular Attention Paid to Proportion and Boltzmann Selection.In[50],1993:124~131
[35]刘长良,于希宁,姚万业等.基于遗传算法的火电厂热工过程模型辨识.中国电机工程学报,2003,23(3):170~174
[36]焦嵩鸣,韩璞,黄宇,李永玲.模糊量子遗传算法及其在热工过程模型辨识中的应用.中国电机工程学报,20072,7(5):88~92
[37]白焰,蒋毅恒,朱耀春,董玲,李霄飞.基于遗传编程的火电厂主汽温系统建模研究.系统仿真学报,2008,20(4):1076~1079
[38]章臣樾.锅炉动态特性及其数学模型.北京:水利电力出版社.1987
[39]王娜娜,边小君,曹云娟.600M级超临界与亚临界锅炉技术比较.浙江电力,2003,31(6)34~35
[40]王小平,曹立明.遗传算法理论应用与软件实现.西安:西安交通大学出版社,2002
[41] Eshelman L,Schaffer J.Real-coded genetical gorithms and interval-schema [C],In:Whitley Ded,Foundations of genetic algorithms 2,San Mateo:Morgan Kaufmann Publishers,1992,187~202
[42]周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999
[43]贺尚红,唐华平,段吉安,等.基于遗传算法的连续系统模型辨识.中南工业大学学报(自然科学版),1998,29(3):280~283
[44]方崇智,萧德云.过程辨识.北京:清华大学出版社,1988
[45]李遵基.热工自动控制系统.北京:中国电力出版社,1999,125～140
[46]王恭良,沈炯,李益国,潘蕾.适用于控制的600 MW超临界机组的简化模型.江苏电机工程,2008,27(4):1~7
[47]张法文.热工过程控制系统分析、设计和调试.中国电力出版社,1997
[48]白焰.神经网络控制在钢球磨煤机中间储仓式制粉系统中的应用:[博士学位论文].沈阳:东北大学,1998
[49] Yan Bai,Guotian Yang,and Xingyi Wang.Decoupling Controller Based on Neural Network,Proceeding of International Conference on Neural Information Processing,1994(2):1141~1145
[50] Bai Yan,Chaitianyou. Neural Decouping Control Based on Time-domain Response Curve Learning.Proceeding of Asian Control Conference ,1997,3: 827~830
[51] A V Ooyen and B Nienhuis.Improving the Convergence of the Back Propagation Algorithm.Neural Networks,1992,3(1):95~100
[52]樊泉桂.超临界锅炉中间点温度控制问题分析.锅炉技术,2005,36(6):1~4
[2]陈永阳,林荣文.浅析超临界机组给水系统.电站辅机,2005,4(95):16~18
[3]周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999
[4]雷德明.多维实数编码遗传算法.控制与决策,2000,15(3):239~241
[5]菅倩.基于矩阵编码遗传算法的研究与应用,[学位论文]:太原,太原理工大学,2008.5
[6] Schraudolph N N,Belew R K.Dynamic Parameter Encoding for Genetic Algorithms.Machine Learning,1992,9(1):9~21
[7]唐飞,滕弘飞,刘峻,等.带有能力约束的二维装填布局问题.宇航学报,2000,21(25)0~57
[8]张思才,张方晓.一种遗传算法适应度函数的改进方法.计算机应用与软件,2006,23(2):108~110
[9] S W Mahfoud.Genetic Drift in SharingMethods[C] .0-7803-1899-4/94,1994,IEEE,62~72
[10] V Petridis,SKazarlis.Varying Quality Function in Genetic Algorithms and the Cutting Problem[C].0-7803-1899-4/94,1994,IEEE,166~169
[11]杨旭东,张彤.遗传算法应用于系统在线辨识研究.哈尔滨工业大学学报,2000,32(1):102~105
[12] Back T,Hoffmeister F.Extended Election Mechanisms in Genetic Algorithms. In[49],1991:92~99
[13] Maza M D L,Tidor B.An Analysis of Selection Procedures with Particular Attention Paid to Proportion and Boltzmann Selection.In[50],1993:124~131
[14] D T Pham,G Jin.Genetic Algorithms Using Gradient-like reproduction operator. Electronics Letters,1995,31(18):1558~1559
[15] Michalewicz .ZGenetic Algorithms+Data structure=Evolution Programs[M] Springer-Verlag,Berlin1992
[16] Back T.Selection pressure in evolutionary algorithms:a Characterization of selection mechanisms[C].Proceeding of the First IEEE conference on Evolutionary computation,IEEE press Orlando,FL.1994:57~62
[17]邝航宇,金晶,苏勇.自适应遗传算法交叉变异算子的改进.计算机工程与应用,2006,12:93~96
[18]李良敏,温广瑞,王生昌.遗传算法中遗传操作的改进策略.计算机应用与软件, 2009, 26 (6):27~30
[19]张明辉,王尚锦.自适应搜索的改进遗传算法及其应用.西安交通大学学报,2002,36(3):226~229
[20] Feng Q X,Francesco P.Theoretical Analysis of Evolutionary Algorithms with an Infinite Population Size in Continuous Space Part I and PartⅡ:and Basic Properties of Selection and Mutation.IEEE Trans Neural Networks,1994,5(1):102~129
[21] Pan Z J,Kang L S,Nie S X.Evolving Both the Topology and Weights of Neural Network. Parallel Algorithms and Applications.1996,9:299~307
[22] Xie Nan,Chen Ying-jun.Improving strategies on genetic algorithm and its application to the optimum design of bridges under earthquake.Gong Cheng Li Xue /Engineering Mechanics 173,2000 Tsinghua Univ 31~36
[23] Wang Lei,Jiao Li-cheng.Novel genetic algorithm based on immunity[C]. Proceedings-IEEE International Symposium on Circuits and Systems 5 May 28-May 31 2000 Sponsored by:IEEE Circuits and Systems Society IEEE V385-V388
[24]骆晨钟,邵惠鹤.采用混沌变异的进化算法.控制与决策,2000,15(9):557~560
[25]叶在福,单渊达.基于多种群遗传算法的输电系统扩展规划.电力系统自动化,2000,10(5):24~27
[26]孟佳娜,王立宏.基于异构多种群的一种改进遗传算法.微电子学与计算机,2004,21(6):71~73
[27] Srinivas M,Patnaik L M.Adaptive probabilities of crossover and mutation in genetic algorithms,IEEE Trans on Systems,Man and Cybernetics,1994,24(4):656~667
[28]陈小平,于盛林.实数遗传算法交叉策略的改进.电子学报,2003,31(1):71~74
[29]王凌,黄璇,郑大钟.一类带筛选策略的改进遗传算法及其性能分析.控制与决策,2004,19(11):1290~1293
[30]杨金明,吴捷,钟丹虹.多目标优化问题中一种改进的遗传算法.华南理工大学学报(自然科学版),2001,29(12):65~67
[31]姚俊峰,梅炽,彭小奇.混沌遗传算法(CGA)的应用研究及其优化效率评价[J].自动化学报,2002,28(6):935~942
[32]王正志,薄涛.进化计算.国防科技大学出版社,2000
[33] Back T,Hoffmeister F.Extended Election Mechanismsin Genetic Algorithms. In[49],1991:92~99
[34] Maza M D L,Tidor B.An Analysis of Selection Procedures with Particular Attention Paid to Proportion and Boltzmann Selection.In[50],1993:124~131
[35]刘长良,于希宁,姚万业等.基于遗传算法的火电厂热工过程模型辨识.中国电机工程学报,2003,23(3):170~174
[36]焦嵩鸣,韩璞,黄宇,李永玲.模糊量子遗传算法及其在热工过程模型辨识中的应用.中国电机工程学报,20072,7(5):88~92
[37]白焰,蒋毅恒,朱耀春,董玲,李霄飞.基于遗传编程的火电厂主汽温系统建模研究.系统仿真学报,2008,20(4):1076~1079
[38]章臣樾.锅炉动态特性及其数学模型.北京:水利电力出版社.1987
[39]王娜娜,边小君,曹云娟.600M级超临界与亚临界锅炉技术比较.浙江电力,2003,31(6)34~35
[40]王小平,曹立明.遗传算法理论应用与软件实现.西安:西安交通大学出版社,2002
[41] Eshelman L,Schaffer J.Real-coded genetical gorithms and interval-schema [C],In:Whitley Ded,Foundations of genetic algorithms 2,San Mateo:Morgan Kaufmann Publishers,1992,187~202
[42]周明,孙树栋.遗传算法原理及应用.北京:国防工业出版社,1999
[43]贺尚红,唐华平,段吉安,等.基于遗传算法的连续系统模型辨识.中南工业大学学报(自然科学版),1998,29(3):280~283
[44]方崇智,萧德云.过程辨识.北京:清华大学出版社,1988
[45]李遵基.热工自动控制系统.北京:中国电力出版社,1999,125～140
[46]王恭良,沈炯,李益国,潘蕾.适用于控制的600 MW超临界机组的简化模型.江苏电机工程,2008,27(4):1~7
[47]张法文.热工过程控制系统分析、设计和调试.中国电力出版社,1997
[48]白焰.神经网络控制在钢球磨煤机中间储仓式制粉系统中的应用:[博士学位论文].沈阳:东北大学,1998
[49] Yan Bai,Guotian Yang,and Xingyi Wang.Decoupling Controller Based on Neural Network,Proceeding of International Conference on Neural Information Processing,1994(2):1141~1145
[50] Bai Yan,Chaitianyou. Neural Decouping Control Based on Time-domain Response Curve Learning.Proceeding of Asian Control Conference ,1997,3: 827~830
[51] A V Ooyen and B Nienhuis.Improving the Convergence of the Back Propagation Algorithm.Neural Networks,1992,3(1):95~100
[52]樊泉桂.超临界锅炉中间点温度控制问题分析.锅炉技术,2005,36(6):1~4