.262 北京科技大学学报 第32卷 预测出口带钢厚度 预测出口 预测轧件塑性Q 贝叶斯 带钢厚度 神经网络 过程控制 数据 (C+Q) 计算机 采集装置 定设定h 反馈 △h/C 控制器 调节量 P -1机架预测 前馈 U 控制器 压下装置 轧辊 出口厚度 +】机架 △H O△HIC i机架 图13贝叶斯神经网络综合应用原理图 Fig 13 Integrated application schenatic of the Bayesian Bayesian neural netorks 通过预测带钢出口厚度、带钢的塑性系数,能够很好 [7]WeiD.Zhang M L JiangZ J et al Neural netork non-linear 地实现带钢厚度实时监控和调整压下量,为进一步 modelling based on Bayesian methods Canput Eng Appl 2005. 41(11):5 在线预测和控制带钢厚度提高广阔应用前景. (魏东,张明廉,蒋志坚,等,基于贝叶斯方法的神经网络非线 性模型辨识.计算机工程与应用,200541(11):5) 参考文献 [8]Sun Y K.Model of Contml Hot Strip M ill Beijing Metallgical [1]Zhang JZ Application of force AGC parmeter equation and vari Industry Press 2002 able stiffness m ill Meta ll Ind Au tomn.1984.3(1):24 (孙一康.带钢热连轧的模型与控制北京:冶金工业出版社, (张进之,压力AGC系统参数方程及变刚度轧机分析,冶金自 2002) 动化,1984,3(1):24) [9]Dong C.H.Matlab Neurl Net Application Beijing National [2]Harman D.Meetng the challenges of intelligent system for Defense Industry Press 2005 Dofasco's tandon col mill Iron Steel Eng 1996.45(6):134 (董长虹,Matb神经网络与应用-北京:国防工业出版社, [3]Ormann B Modemization of the automnation in the hot w ide strip 2005) m ill at VoestA bne Stahl Metall Plant Technol Int 1994.6 26 [10]Zhu M.Data Mining Hefei China Science and Technobgy [4]Huang Q Application of neuml nehvork n breakout prediction University Press 2002 systm for continuous casting n Baosteel Baosteel Technol 1999 (朱明.数据挖掘.合肥:中国科学技术大学出版社,2002) (1):40 [11]Ticu A K.Jiang Z Y,Lu C A 3D finite element analysis of the (黄棋·神经网络在宝钢连铸漏钢预报系统中的应用,宝钢技 hot molling of strip with lubrication J Mater Process Technol 术,1999(1):40) 2002125/126.638 [5]HajeerM N.Basheer IA.A hybril Bayesian euml netork ap- [12]Jia T Liu Z Y.Hu H F et al Mechanical property prediction proach for pmobabilistic modeling of bacterial gwh/no gmwh in- for hot molled SPA H steel usng Bayesian neural network J terface Int J Food M icmobiol 2003 82:233 Norheastem Univ Nat Sci 2008 29(4):521 [6]Foresee F D.Hagan M T.Gauss New ton approxination to Bayes- (贾涛,刘振宇,胡恒法,等.基于贝叶斯神经网络的SPA一H an leaming Pmceedings of the Intemational Confernce on Neuml 热轧板力学性能预测,东北大学学报:自然科学版,200829 Networks Houston 2007 (4):521)北 京 科 技 大 学 学 报 第 32卷 图 13 贝叶斯神经网络综合应用原理图 Fig.13 IntegratedapplicationschematicoftheBayesianBayesianneuralnetworks 通过预测带钢出口厚度、带钢的塑性系数能够很好 地实现带钢厚度实时监控和调整压下量为进一步 在线预测和控制带钢厚度提高广阔应用前景. 参 考 文 献 [1] ZhangJZ.Applicationofforce-AGCparameterequationandvari- ablestiffnessmill.MetallIndAutom19843(1):24 (张进之.压力 AGC系统参数方程及变刚度轧机分析.冶金自 动化19843(1):24) [2] HartmanD. Meetingthechallengesofintelligentsystem for Dofascoʾstandomcoldmill.IronSteelEng199645(6):134 [3] OrtmannB.Modernizationoftheautomationinthehotwidestrip millatVoest-AlpineStahl.MetallPlantTechnolInt19946:26 [4] HuangQ.Applicationofneuralnetworkinbreakoutprediction systemforcontinuouscastinginBaosteel.BaosteelTechnol1999 (1):40 (黄棋.神经网络在宝钢连铸漏钢预报系统中的应用.宝钢技 术1999(1):40) [5] HajmeerMNBasheerIA.AhybridBayesian-neuralnetworkap- proachforprobabilisticmodelingofbacterialgrowth/no-growthin- terface.IntJFoodMicrobiol200382:233 [6] ForeseeFDHaganMT.Gauss-NewtonapproximationtoBayes- ianlearning∥ProceedingsoftheInternationalConferenceonNeural Networks.Houston2007 [7] WeiDZhangMLJiangZJetal.Neuralnetworknon-linear modellingbasedonBayesianmethods.ComputEngAppl2005 41(11):5 (魏东张明廉蒋志坚等.基于贝叶斯方法的神经网络非线 性模型辨识.计算机工程与应用200541(11):5) [8] SunYK.ModelofControlHotStripMill.Beijing:Metallurgical IndustryPress2002 (孙一康.带钢热连轧的模型与控制.北京:冶金工业出版社 2002) [9] DongCH.MatlabNeuralNet& Application.Beijing:National DefenseIndustryPress2005 (董长虹.Matlab神经网络与应用.北京:国防工业出版社 2005) [10] ZhuM.DataMining.Hefei:ChinaScienceandTechnology UniversityPress2002 (朱明.数据挖掘.合肥:中国科学技术大学出版社2002) [11] TieuAKJiangZYLuC.A3Dfiniteelementanalysisofthe hotrollingofstripwithlubrication.JMaterProcessTechnol 2002125/126:638 [12] JiaTLiuZYHuHFetal.Mechanicalpropertyprediction forhotrolledSPA-H steelusingBayesianneuralnetwork.J NortheasternUnivNatSci200829(4):521 (贾涛刘振宇胡恒法等.基于贝叶斯神经网络的 SPA--H 热轧板力学性能预测.东北大学学报:自然科学版200829 (4):521) ·262·