基于模糊神经网络的磁浮车悬浮控制器

作者：孙荣斌，王鑫，夏加宽 日期：2008-02-25/span> 浏览：3896 查看PDF文档

基于模糊神经网络的磁浮车悬浮控制器

孙荣斌，王鑫，夏加宽
（沈阳工业大学 电气工程学院，辽宁 沈阳 110023）

摘要：以EMS型磁悬浮车单电磁铁为研究对象，建立了系统的数学模型。为增强悬浮系统的抗干扰能力和负载能力，针对其非线性特性，采用神经网络和模糊控制理论相结合的方法设计了非线性控制器FNNC。该控制器继承了模糊控制鲁棒性强以及神经网络的非线性逼近和自学习能力，运用加速反传训练方法对参数进行调整，并通过TMS320F2812型DSP实现单点悬浮用。实验及仿真表明，该控制器使系统悬浮稳定,动态性能好，与传统的PID控制器相比具有更好的快速跟踪性能和抗干扰能力。
关键词：磁悬浮列车；非线性；模糊神经网络；数字信号处理器
中图分类号：TP273+.5文献标识码：A文章编号：1001-4551(2008)01-0086-04

Suspension system of maglev train based on fuzzy neural network controller
SUN Rongbin, WANG Xin, XIA Jiakuan
(College of Electrical Engineering, Shenyang University of Technology, Shenyang 110023, China)
Abstract: The single electromagnet of EMS maglev train was chosen to be the controlled object. The nonlinear mathematic model of electromagnetic suspension system was derived. In order to improve the ability of disturbance rejection and load, considering the nonlinear characteristic, an nonlinear controller FNNC was designed which combined the neural network and the fuzzy control theory. FNNC inherited the robust characteristic of fuzzy control and the strong ability of nonlinear processing and selflearning of neural network. The training method of accelerating backpropagation algorithm was used to adjust the parameter. Singlepoint suspension was realized through TMS3200F2812 DSP. The test and emulation results show that the suspension point is stable and the performance of dynamic is good. FNNC is preferable to PID controller in respect of fast tracking performance and disturbance rejection performance.
Key words: maglev train; nonlinear; fuzzy neural network; digital signal processor (DSP)
参考文献(Reference):
［1］张东升,梅雪松.一种磁悬浮系统的降阶方法［J］.系统仿真学报,2007,19（10）：2256-2259.
［2］MIYAZAKI K, YANG Z J, KANAE S, et al. Adaptive robust nonlinear control of magnetic levitation system via DSC technique［J］. Electrical Engineering in Japan,2004,149（4）:42-51.
［3］曲萍萍，魏绍义.磁悬浮系统的模糊控制器设计［J］.控制工程，2006，14（z1）：78-80.
［4］罗成，李云钢.磁浮列车悬浮系统的神经网络建模研究［J］.计算机仿真，2006，23（1）：144-146.
［5］SUN C T. Rulebase structure identification in an adaptive network based fuzzy inference system［J］. IEEE Transactions on Fuzzy Systems,1994，2(1)：64-73.
［6］张智星，孙春在.神经模糊和软计算［M］.西安：西安交通大学出版社，2000.
［7］郑明山.模糊神经网络控制器算法的研究［J］.中国水运，2007,5（3）：114-115.