水泥篦冷机出口熟料温度自适应辨识模型
|
摘要
为了优化控制系统,建立篦冷机温度熟料出口的识别模型,利用篦冷机内熟料换热机理,找出熟料冷却过程的关键影响因素;利用回声状态网络辨识篦冷机运行数据,基于递归最小二乘法推导网络的在线学习算法,实现权值自适应调整,从而建立了篦冷机出口熟料温度的自适应辨识模型.仿真实验可知,在系统发生变化时构建的模型能够自适应调整网络输出权值矩阵,使模型快速收敛.与其他离线方法相比,提出的熟料出口温度的自适应模型更加持久有效,可以作为辨识模型指导篦冷机的控制.
In order to optimize the control system and establish the identification model of the clinker outlet of the grate cooler temperature, the key influencing factors of the clinker cooling process are found out according to the heat exchange mechanism of the clinker in the grate cooler. The operating data of grate cooler is trained by echo state network(ESN).Based on the recursive least squares(RLS), the online learning algorithm of the network is derived to realize the adaptive adjustment of the weights in ESN. Then the adaptive model of the clinker temperature in the grate cooler is obtained.Simulation experiments show that when the system changes, the built model can quickly converge by adaptively adjusting the output weight matrix of the network. Compared with other off-line methods, the proposed adaptive model of the clinker temperature is more prolonged and effective. Consequently, it can be used as an identification model to guide the control to the grate cooler so as to improve the efficiency of grate coolers.
In order to optimize the control system and establish the identification model of the clinker outlet of the grate cooler temperature, the key influencing factors of the clinker cooling process are found out according to the heat exchange mechanism of the clinker in the grate cooler. The operating data of grate cooler is trained by echo state network(ESN).Based on the recursive least squares(RLS), the online learning algorithm of the network is derived to realize the adaptive adjustment of the weights in ESN. Then the adaptive model of the clinker temperature in the grate cooler is obtained.Simulation experiments show that when the system changes, the built model can quickly converge by adaptively adjusting the output weight matrix of the network. Compared with other off-line methods, the proposed adaptive model of the clinker temperature is more prolonged and effective. Consequently, it can be used as an identification model to guide the control to the grate cooler so as to improve the efficiency of grate coolers.
引文
[1]MUJUMDAR K S,GANESH K,KULKARNI S B,et al.Rotary cement kiln simulator(RoCKS):Integrated modeling of pre-heater,calciner,kiln and clinker cooler.Chemical Engineering Science,2007,62(9):2590-2607.
[2]JU Ruixi.Causes and solutions for the production of“Red River”by the grate cooler.Cement,2014(9):34.(琚瑞喜.篦冷机产生“红河”的原因及解决措施.水泥,2014,2014(9):34.)
[3]AHMAD R,KHAN T A,AGARWAL V.Mass and energy balance in grate cooler of cement plant.International Journal of Scientific Engineering and Technology,2013,2(7):631-637.
[4]LIU Z,WANG Z,YUAN M,et al.Thermal efficiency modelling of the cement clinker manufacturing process.Journal of the Energy Institute,2015,88(1):76-86.
[5]WANG Meiqi,LIU Bin,WEN Yan,et al.Temperature soft-sensing of cement clinker in grate cooler.Journal of Mechanical Engineering,2016,52(6):159-165.(王美琪,刘彬,闻岩,等.篦冷机内水泥熟料温度的软测量.机械工程学报,2016,52(6):159-165.)
[6]LI Haibin,HAO Xiaochen,LIU Bin,et al.Modelling for clinker flow of the grate cooler based on fuzzy neural networks.Chinese Journal of Scientific Instrument,2005,S2(8):171-173.(李海滨,郝晓辰,刘彬,等.基于模糊神经网络的篦式冷却机熟料流动建模.仪器仪表学报,2005,S2(8):171-173.)
[7]CAI Ning,YAN Wenjun.Application of RBF predictive control in cement grate cooling system.Cement,2010,2010(10):58-60.(蔡宁,颜文俊.RBF预测控制在水泥篦冷机系统中的应用.水泥,2010,(10):58-60.)
[8]LI Changxian,ZHU Jin.Application study of single input and dual output fuzzy-predictive coordination control on grate cooler.Journal of the Chinese Ceramic society,2002,30(6):707-711.(李常贤,诸静.单入双出的模糊预测协调控制在篦冷机系统中的应用研究.硅酸盐学报,2002,30(6):707-711.)
[9]KARIKOV,RUBANOV V,KLASSEN V.Construction of a dynamic neural network model as a stage of grate cooler automation.World Applied Sciences Journal,2013,25(2):227-232.
[10]SCHNEIDER M.Process technology for efficient and sustainable cement production.Cement Concrete Research,2015,78:14-23.
[11]WANG Meiqi,LIU Bin,WEN Yan,et al.Seepage heat transfer model and numerical simulation of grate cooler.Journal Of Mechanical Engineering,2014,50(24):152-157.(王美琪,刘彬,闻岩,等.篦冷机的渗流传热模型与数值模拟.机械工程学报,2014,50(24):152-157.)
[12]WEN Y,WANG J,YUE H,et al.Experimental study on the permeability and thermal resistance of the cement clinker accumulation.Heat Transfer―Asian Research,2016,45(2):134-143.
[13]SHAO W,CUI Z,WANG N H,et al.Numerical simulation of heat transfer process in cement grate cooler based on dynamic mesh technique.Science China Technological Sciences,2016,59(7):1065-1070.
[14]FENG Shaohang,XU Delong,LI Hui,et al.Simulation study on heat transfer between clinker and gas of the grate cooler.Journal of Xi’an University of Architecture&Technology(Natural Science Edition),2007,39(2):224-229.(冯绍航,徐德龙,李辉,等.篦冷机中气固两相换热过程的模拟研究.西安建筑科技大学学报(自然科学版),2007,39(2):224-229.)
[15]LIU B,WANG M,WEN Y,et al.Research on control mechanism model of grate cooler based on seepage heat transfer theory.Journal of Computational Information Systems,2013,20(9):8281-8288.
[16]HWANG C L,JAN C.Recurrent-neural-network-based multivariable adaptive control for a class of nonlinear dynamic systems with timevarying delay.IEEE Transactions on Neural Networks and Learning Systems,2016,27(2):388-401.
[17]RODAN A,TINO P.Minimum complexity echo state network.IEEETransactions on Neural Networks,2011,22(1):131-144.
[18]JAEGER H.The“echo state”approach to analysing and training recurrent neural networks-with an erratum note.German National Research Center for Information Technology GMD Technical Report,2001,148(34):13.
[19]JAEGER H.Short term memory in echo state networks.Bonn,Germany:GMD-Forschungszentrum Informationstechnik,2001.
[20]JAEGER H.Tutorial on training recurrent neural networks,covering BPPT,RTRL,EKF and the“echo state network”approach.Bonn:GMD-Forschungszentrum Informationstechnik,2002.
[21]HERTZBERG J,JAEGER H,SCHONHERR F.Learning to ground fact symbols in behavior-based robots.Proceedings of the 15th European Conference on Artificial Intelligence.Netherlands:IOS Press,2002:708-712.
[22]BADONI M,SINGH A,SINGH B.Variable forgetting factor recursive least square control algorithm for DSTATCOM.IEEE Transactions on Power Delivery,2015,30(5):2353-2361.
[23]FARHANG B.Adaptive Filters:Theory and Applications.New Jersey:John Wiley&Sons,2013:417-431.
[2]JU Ruixi.Causes and solutions for the production of“Red River”by the grate cooler.Cement,2014(9):34.(琚瑞喜.篦冷机产生“红河”的原因及解决措施.水泥,2014,2014(9):34.)
[3]AHMAD R,KHAN T A,AGARWAL V.Mass and energy balance in grate cooler of cement plant.International Journal of Scientific Engineering and Technology,2013,2(7):631-637.
[4]LIU Z,WANG Z,YUAN M,et al.Thermal efficiency modelling of the cement clinker manufacturing process.Journal of the Energy Institute,2015,88(1):76-86.
[5]WANG Meiqi,LIU Bin,WEN Yan,et al.Temperature soft-sensing of cement clinker in grate cooler.Journal of Mechanical Engineering,2016,52(6):159-165.(王美琪,刘彬,闻岩,等.篦冷机内水泥熟料温度的软测量.机械工程学报,2016,52(6):159-165.)
[6]LI Haibin,HAO Xiaochen,LIU Bin,et al.Modelling for clinker flow of the grate cooler based on fuzzy neural networks.Chinese Journal of Scientific Instrument,2005,S2(8):171-173.(李海滨,郝晓辰,刘彬,等.基于模糊神经网络的篦式冷却机熟料流动建模.仪器仪表学报,2005,S2(8):171-173.)
[7]CAI Ning,YAN Wenjun.Application of RBF predictive control in cement grate cooling system.Cement,2010,2010(10):58-60.(蔡宁,颜文俊.RBF预测控制在水泥篦冷机系统中的应用.水泥,2010,(10):58-60.)
[8]LI Changxian,ZHU Jin.Application study of single input and dual output fuzzy-predictive coordination control on grate cooler.Journal of the Chinese Ceramic society,2002,30(6):707-711.(李常贤,诸静.单入双出的模糊预测协调控制在篦冷机系统中的应用研究.硅酸盐学报,2002,30(6):707-711.)
[9]KARIKOV,RUBANOV V,KLASSEN V.Construction of a dynamic neural network model as a stage of grate cooler automation.World Applied Sciences Journal,2013,25(2):227-232.
[10]SCHNEIDER M.Process technology for efficient and sustainable cement production.Cement Concrete Research,2015,78:14-23.
[11]WANG Meiqi,LIU Bin,WEN Yan,et al.Seepage heat transfer model and numerical simulation of grate cooler.Journal Of Mechanical Engineering,2014,50(24):152-157.(王美琪,刘彬,闻岩,等.篦冷机的渗流传热模型与数值模拟.机械工程学报,2014,50(24):152-157.)
[12]WEN Y,WANG J,YUE H,et al.Experimental study on the permeability and thermal resistance of the cement clinker accumulation.Heat Transfer―Asian Research,2016,45(2):134-143.
[13]SHAO W,CUI Z,WANG N H,et al.Numerical simulation of heat transfer process in cement grate cooler based on dynamic mesh technique.Science China Technological Sciences,2016,59(7):1065-1070.
[14]FENG Shaohang,XU Delong,LI Hui,et al.Simulation study on heat transfer between clinker and gas of the grate cooler.Journal of Xi’an University of Architecture&Technology(Natural Science Edition),2007,39(2):224-229.(冯绍航,徐德龙,李辉,等.篦冷机中气固两相换热过程的模拟研究.西安建筑科技大学学报(自然科学版),2007,39(2):224-229.)
[15]LIU B,WANG M,WEN Y,et al.Research on control mechanism model of grate cooler based on seepage heat transfer theory.Journal of Computational Information Systems,2013,20(9):8281-8288.
[16]HWANG C L,JAN C.Recurrent-neural-network-based multivariable adaptive control for a class of nonlinear dynamic systems with timevarying delay.IEEE Transactions on Neural Networks and Learning Systems,2016,27(2):388-401.
[17]RODAN A,TINO P.Minimum complexity echo state network.IEEETransactions on Neural Networks,2011,22(1):131-144.
[18]JAEGER H.The“echo state”approach to analysing and training recurrent neural networks-with an erratum note.German National Research Center for Information Technology GMD Technical Report,2001,148(34):13.
[19]JAEGER H.Short term memory in echo state networks.Bonn,Germany:GMD-Forschungszentrum Informationstechnik,2001.
[20]JAEGER H.Tutorial on training recurrent neural networks,covering BPPT,RTRL,EKF and the“echo state network”approach.Bonn:GMD-Forschungszentrum Informationstechnik,2002.
[21]HERTZBERG J,JAEGER H,SCHONHERR F.Learning to ground fact symbols in behavior-based robots.Proceedings of the 15th European Conference on Artificial Intelligence.Netherlands:IOS Press,2002:708-712.
[22]BADONI M,SINGH A,SINGH B.Variable forgetting factor recursive least square control algorithm for DSTATCOM.IEEE Transactions on Power Delivery,2015,30(5):2353-2361.
[23]FARHANG B.Adaptive Filters:Theory and Applications.New Jersey:John Wiley&Sons,2013:417-431.