Journal of South China University of Technology (Natural Science Edition) ›› 2005, Vol. 33 ›› Issue (12): 75-77,91.

Previous Articles     Next Articles

Semi-Active Suspension System of Train Based on Nonlinear Neural Networks

Ding Wen-si1  Bu Ji-ling2   

  1. 1.College of Automotive Engineering,South China Univ.of Tech.,Guangzhou 5 10640,Guangdong,China;2.Vehicle Research Institute,Southwest Jiaotong Univ.,Chengdu 610031,Sichuan,China
  • Received:2005-01-14 Online:2005-12-25 Published:2005-12-25
  • Contact: 丁问司(1968-),男,副教授,主要从事轨道车辆悬架及电液控制研究 E-mail:traindamper@tom.com
  • About author:丁问司(1968-),男,副教授,主要从事轨道车辆悬架及电液控制研究
  • Supported by:

    广东省自然科学基金博士科研启动基金资助项目(04300105);华南理工大学自然科学基金资助项目(E504 1010)

Abstract:

In order to improve the lateral ride quality of train without endangering safety,this paper illustrates a lat-eral semi-active suspension system based on the skyhook rule.A controlling model of the proposed system is then put forward based on n0nlinear neural networks,with a neural identifier and a controller being also designed.The simulated and tested resuhs show that.compared with the status without any controlling step,the peak of lateral ac-celerati0n of the semi-active suspension system controlled by neural networks decreases obviously,and that the lateral ride quality of train is efectively improved,with the lateral ride item (the value of )being improved by 9.04% in the simulation and 15.9% in the actual testing.

Key words: train, semi-active suspension, damper, nonlinear neural network