摘要
数学模型在现代的工业过程中十分的重要,是一系列控制理论应用的前提,是优化控制系统的基础。于是,实际中对数学模型精确度的要求越来越高,相对于机理分析法来建模,辨识有着自己的优势,由于非线性普遍的存在于实际生活之中,怎么建立非线性部分的模型对于工程实践来说有着至关重要的作用。本文重点介绍了模块化非线性模型中Hammerstein和Wiener模型以及它们的辨识方法,并通过对例子进行仿真,得到辨识结果。本文主要做了以下几个方面的工作：
1、介绍了系统辨识的一些基础知识和几种经典的辨识方法,描述了一止匕Hammerstein模型以及它们的工业应用,针对了一种火花点火发动机的Hammerstein模型进行仿真,说明了递推最小二乘可以应用在Hammerstein模型的辨识问题上,还仿真了几种Hammerstein模型,说明了基于辅助模型的递推最小二乘算法在处理这类问题上有很好的辨识效果；
2、描述了粒子群算法(PSO)及它的一些改进算法,用仿真例子说明了PSO以及基于全局收敛的改进PSO算法即SPSO算法在Hammerstein模型上可以成功的进行应用并得到好的参数估计。
3、针对一种可以统一描述各种非线性特性的非线性模型进行辨识方法研究,并成功的使用PSO算法得到了拥有这种统一非线性部分的Hammerstein模型的参数估计,说明了通过对其中参数的选择可以处理9种不同非线性特性Hammerstein模型的辨识问题。
4、使用SPSO算法来处理一种Wiener模型的辨识问题,用仿真例子说明了在有噪声干扰的情况下SPSO也能获得良好的参数估计,最后提出了一种基于SPSO算法的2步辨识法来辨识Wiener模型,并通过例子说明了该算法的可行性。
Mathematical model is very important for modern idustrial engineering,itis the foundation of a series of control theory and system optimizationcontrol.Then,the precision requirements of mathematical model is more andmore taller in the actual.Relative to the mechanism analysis which can be useto modeling,the identification has its own advantages.Because of the commonexists in nature,how to establish the model of the nonliear system is a vitalrole of engineering practicd. This paper introduces the nonliner block orientedmodel include Hammerstein and Wiener model and their identificationmethods.By simulating some example,getitng identificaiton results.This papermain consist of the following several aspects work.
1Introduceing some basic knowledge of system identification andseveral kinds of classic identification methods,described some Hammersteinmodel and its industrial application,in the view of a spsrk ignition enginesHammerstein model of simulation,explain recursive least squares can beapplied in identification problem of Hammerstrin model.Simulating several Hammerstein model and compared identification resules of the auxiliarymodel recursive least suqare algorithm and a first filtering of least squaremehod.
2Describing the paricle swarm optimization(PSO) algorithm and itsmodified algorithm.Simluating some examples shows that the PSO andmodified PSO based on global convergence which called SPSO algorithm cansuccessfully in applied to identification problem of the Hammerstein modeland get good parameter estimation.
3According to a unified model which can description of variousnonlinear characteristics of the nonlinear model and researching itsidentificaion methods.Successfully estimating parameters of Hammersteinmodel with the unified nonlinear model by PSO algorithm.Through theselection of parameters of the unified model can deal with9differentnonlinear characteristics of Hammerstein model identification problem.
4Using SPSO algorithms deal with a Wiener model identificationproblem.The simulation shows that SPSO also can get well parametersestimation under a noise interference.Finally puts forward a identificationmethod of two steps to identify the Wiener model based on SPSOalgorithm.The simulation shows that the algorithm is feasible.
引文
[1]C. Andrieu, A. Doucet, S.S. Singh, V.B. Tadic, Particle methods for change detection, system identification, and control, Proceedings of the IEEE92(2004)423-438.
[2]方崇智,萧德云.过程辨识[M].北京：清华大学出版社,1988.
[3]朱豫才·著.张湘平,虞水俊,孙志强,胡德文·译.过程控制的多变量系统辨识[M].长沙：国防科技大学出版社,2004.
[4]吴广玉.系统辨识与自适应[M].哈尔滨：哈尔滨工业大学出版社,1987.
[5]刘福才.非线性系统的模糊模型辨识及其应用[M].北京：国防工业出版社,2006.
[6]Zadeh L. A. From Circuit Theory to System Theory[J]. Proc. IRE,1962,50(5):856-865.
[7]Eykhoff P. System Identification-Parameter and State Estimation[M] John Wiley&Sons,1974.
[8]Ljung L.Convergence Analysis of Parametric Identification Methods[J].IEEE Trans.on Automatic Control,1978,AC-23:770-783.
[9]Ljung L. Asymptotic variance expressions for identified black-box transfer function models [J].IEEE Transactions on Automatic Control,1985,AC-30:834-844.
[10]张承乾,张国强.系统辨识与参数估计.北京：机械工业出版社,1986.
[11]Scherzen. M. The Volterra and Wiener Theories of Nonlinear Systems. New York, Wiley,1980.
[12]Suleimam W,Monin A. New method for identifying finite degree Volterra series [J].Automatica,2008,44(2):488-497.
[13]金哲,宋执环,何加铭.基于简化volterra级数的射频功率放大器建模与辨识[J].电路与系统学报,2008,13(5)：90-94.
[14]王文正,何开锋,欧文,韩崇昭.基于volterra泛函级数的非线性系统的鲁棒辨识[J].控制理论与应用,2003,20(3)：432-436.
[15]欧文,韩崇昭,王文正volterra泛函级数在非线性系统辨识中的应用[J].控制与决策,2002,17(2)：239-242.
[16]魏瑞轩,韩崇昭.一种基于Volterra模型的非线性系统具有全解藕结构的递阶式自适应辨识方法[J].控制理论与应用,2002,19(2)：313-316.
[17]朱全民.非线性系统辨识[J].控制理论与应用,1994,H(6)：641-652.
[18]Billings. S. A and Fakhours. S. Y. Identification of Nonlinear System Using the Wiener Model. Electronics Letters,1977,13:502-504.
[19]Narendra. K. S, Gallman. P. G. An Iterative Method for identification of Nonlinear System Using a Hammerstein Model. IEEE Trans,1966, AC-14:546-550.
[20]NARENDRA. K,GALLMAN. P. An iterative method for the identification of nonlinear systems using a Hammerstein model. IEEE Trans on Automatic Control,1966,11(3):546-550.
[21]Narendra K S,Gallman P G An iterative method for the identification of nonlinear systems using a Hammerstein model[J]. IEEE Transactions on Automatic Control,1966, AC-11:546-550.
[22]VorosJ. Parameter identifieation of diseontinuous Hammerstein system[J]. Automatica1997,33(6):1141-1146.
[23]ChangF,LuusR. A noniterative method for identifieation using Hammerstein model[J].IEEE Transactions on Automatic Control,1971, AC-6:464-468.
[24]Bai E W.Identification of systems with hard input nonlinearitis[C].In PersPectives in Control, MoheimaniR,Ed.NewYork:Springer Verlag,2001.
[25]Hjalmarsson H,Gunnarsson S,Gevers M.Optimality and suboptimality of iterative identification and control design seheme[C].In Proeeedings of the American Control Conferenee,1995,4:2559-2563.
[26]Raieh R, Zhou G T, Viberg M.SubsPace based approaches for Wiener system identification[J].IEEE Transactions on Automatic Control,2005,50(10):1629-1634.
[27]Greblicki W. Nonparametric identification of Wiener system[J]. IEEE Transactions on Information Theory,1992,38(10):1487-1493.
[28]Bai E w,Reyland Jr J. Towards identification of Wiener the laest amount of a priori information on the nonlinearity[J]. Automatica,2008,44(4):910-919.
[29]张广莹,邓正隆.基于小波分析的非线性系统Wiener模型辨识[J].电机与控制学报,2001,5(2)：95-97,128.
[30]ANORLD,C.R.and NARENDRA,K.S.. The Characterization and Identification of Systems [R]. Technical Report471, Harvard University.1965.
[31]ALCKSANDROVSKII, N. M. and DEICH,A.M.. Determination of Dynamic Characteristic of Nonlinear Objects [J].Automat. Remote Control,1968,(29):142-160.
[32]HUNG,G and STARK,L. The Kernel Identification Method-Review of Theory, Calculation, Application and Interpretation [J].Math Biosci.,1977,(37):135-170.
[33]BILLINGS,S.A. Identification of Nonlinear System-A Survey. Proc[J]. IEE.1980,127(6):272-285.
[34]李茶玲,孙德保.遗传算法在系统辨识中的应用[J].华中理工大学学报,1998,26(7)：57-58,2005.
[35]边润强,陈增强,袁著祉.一种改进的遗传算法及其在系统辨识中的应用[J].控制与决策,2000,15(5)：623-625.
[36]徐洪泽.一种具有变异性杂交的遗传算法及其在系统辨识中的应用[J].哈尔滨工业大学学报,1997,29(4)：74-75.
[37]BILLINGS,S.A. and etc. Properties of Neural Networks with Applications to Modeling Nonlinear Dynamical Systems[J]. Int. J. Control.1992,55(1):193-224.
[38]邵青,冯汝鹏.非线性系统模糊辨识的新方法[J].控制与决策,2001,16(1)：83-85.
[39]刘福才,关新平,裴润.基于一种新模糊模型的非线性系统模糊辨识[J].控制理论与应用,2003,20(1)：113-116.
[40]张平安,熊学健,李人厚.基于拟非线性模糊模型的复杂系统模糊辨识[J].控制理论与应用,1998,15(2)：286-290.
[41]Ljung L.Soderstrom T. Theroy and Practice of Recursive Identification. MIT Press,Cambridge, Massachusetts,London,England.1983.
[42]艾克霍夫P.著,系统辨识—参数和状态估计.潘科炎,张永光等译.北京：科学出版社,1980.
[43]夏天长.系统辨识—最小二乘法[M].北京：清华大学出版社,1983.
[44]Ding F, Chen T. Identification of dual-rate systems based on finite impulse responsemodels[J]. International Journal of Adaptive Control and Signal Processing2004,18(7):589-598.
[45] Ding F Chen T Combined parameter and output estimation of dual-rate systemsusing an auxiliary model[J].Automatica200440(10):1739-1748.
[46] Ding F Chen T Parameter estimation of dual-rate stochastic systems by using anoutput error method[J] IEEE Transactions on Automatic Control200550(9):1436-1441.
[47] Ding F,Liu X P,Liu G J.Gradient based and leastsquares based iterative identificationmethods for OE and OEMA systems[J] Digital Signal Processing2010,20(3):664-677.
[48] L.Guzzella, C.H.Onder, Introduction to Modeling and Control of Internal CombustionEngine Systems,Springer,Berlin,NewYork,2004.
[49] T.Kara, I. Eker,Nonlinear modeling and identification of a DC motor for bidirectionaloperation with real time experiments,Energy Conversion and Management45(2004)1087–1106.
[50] Kennedy, J. and Eberhart, R. C.(1995). Particle swarm optimization. Proc. IEEE Int'l.Conf. on Neural Networks, IV,1942–1948. Piscataway, NJ: IEEE Service Center.
[51] Shi, Y. and Eberhart, R. C.(1998b). A modified particle swarm optimizer. Proceedingsof the IEEE International Conference on Evolutionary Computation,69-73.Piscataway, NJ: IEEE Press.
[52] Clerc, M.(1999). The swarm and the queen: towards a deterministic and adaptiveparticle swarm optimization. Proc.1999Congress on Evolutionary Computation,Washington, DC, pp1951-1957. Piscataway, NJ: IEEE Service Center.
[53] Shi, Y. and Eberhart, R. C. Fuzzy adaptive particle swarm optimization.2001. Seoul,Korea, IEEE Service Center. Proc. Congress on Evolutionary Computation2001.
[54] V r sJ.“Parameter identification of Wiener systems with discontinuousnonlinearities,” Systems&Control Letters, vol.44, no.5, pp.363–372,2001.
[55] V r sJ.“Recursive Identification of Hammerstein Systems With DiscontinuousNonlinearities Containing Dead-Zones,” IEEE Transactions on Automatic Control,vol.48, no.12, pp.2203–2206,2003.
[56] A. Chaari, S. Rejeb, F. Ben Hmida, H. Messaoud, and M. Gossa,“Iterativeidentification of Wiener model using hysteresis memory-less nonlinearity,”International Journal of Sciences and Techniques of Automatic control&computerengineering IJ-STA, vol.2,no.1, pp.430–441,2008.
[57] Kalafatis,A.,Arifin,N.,Wang, L.,&Cluett,W.R.(1995). A new approach toidentification of pH process based on the Wiener model. Chemical EngineeringScience,50,3693–3701.
[58] Singh,M.,Elloy,J.,Mezencev,R.,&Munro, N.(1978). Applied industrial control.Oxford:Pergamon Press.
[59] Pearson, R. K.,&Pottmann, M.(2000). Gray-box identification of block-orientednonlinear models. Journal of Process Control,10,301–315.
[60] Gilabet, P. L., Montoro, G.,&Bertran, E.(2005). On the Wiener and Hammersteinmodels for power amplifier predistortion. In Proceedings of the Asia–Pacificmicrowave conference (Vol.2, pp.1191–1194).
[61] Bai, E. W.(2002). A blind approach to the Hammerstein–Wiener modelidentification. Automatica,38,967–979.