自适应神经PID在辊筒温控系统中的应用Self-Adaptive Neural PID and Its Application to Temperature Control System of Rollers
张小明,瞿金平
摘要(Abstract)：
针对压延机辊筒三段温度控制以及常规PID控制在非线性的、时变系统中控制效果的局限性,提出了一种基于BP神经网络整定的PID控制方法,同时给出了计算机控制系统设计。由于神经网络具有强大的非线性映射能力及自学习、自适应等优势,通过对系统性能的学习来实现具有最佳组合的PID控制,建立比例、积分和微分3种参数自学习的PID控制器。对压延机辊筒的温度控制试验结果表明,用该方法整定的PID控制系统,逼近精度高、适应性好。
关键词(KeyWords)： 压延机;辊筒;温度控制;BP神经网络;自适应PID控制
基金项目(Foundation): 国家自然科学基金项目(10472034,10590351)
作者(Author): 张小明,瞿金平
参考文献(References)：
- 1 Bajcinca N.Design of robust PID controllers using decou-pling at singular frequencies.Automatica,2006,42(11):1 943~1 949
- 2 Bai R L,Wang J,Wang L F,et al.Design of NN-PID con-troller based on PSO and its FPGA implementation.Dynam-ics of Continuous Discrete and Impulsive Systems-Series A-Mathematical Analysis,2006,13:1 307~1 314
- 3 D′Emilia G,Marra A,Natale E.Use of neural networks forquick and accurate auto-tuning of PID controller.Roboticsand Computer-Integrated Manufacturing,2007,23(2):170~179
- 4 Hovda E,Irgens F,Olsen JE.Non-linear modeling of paperin a hard-nip calender.Atlanta:Tappi Press,1999.153~160
- 5 Gutzeit O,Hanisch HM.Modular modeling of roll-basedcalender manufacturing environments.New York:IEEE,2003.321~328
- 6 Wagner U.A new calender concept for the next century.Atlanta:Tappl Press,1997.111~115
- 7 Fu L D,Shen Q C,Chen K S,et al.The intelligent controlmodels for hydraulic oil temperature.Dynamics of Continu-ous Discrete and Impulsive Systems-Series A-MathematicalAnalysis,2006,13:1 729~1 736
- 8 Bretas AS,Pires L,Moreto M,et al.A BP neural networkbased technique for HIF detection and location on distribu-tion systems with distributed generation.Berlin:Springer-werlag Berlin,2006.608~613
- 9 Fuh K H,Wang S B.Force modeling and forecasting increep feed grinding using improved BP neural network.In-ternational Journal of Machine Tools&Manufacture,1997,37(8):1 167~1 178
- 10 Li Z Y,Deng X M.Uncertainty relation suited to overfit-ting of BP neural network.Journal of Infrared and Millime-ter Waves,2000,19(2):142~144
- 11 Lei L,Wang H J,Cheng Y F.An improved BP algorithmbased on global revision factor and its application to PIDcontrol.Berlin:Springer-Verlag Berlin,2006.1 019~1 026
- 13 Zhang Y,Feng C,Li B L.PID control of nonlinear motor-mechanism coupling system using artificial neural network.Berlin:Springer-Verlag Berlin,2006.1 096~1 103
- 13 Chao C Y H,Hu J S.Development of an enthalpy and car-bon dioxide based demand control ventilation for Indoor airquality and energy saving with neural network control.In-door and Built Environment,2004,13(6):463~475