基于改进Elman神经网络的悬架试验系统  

Suspension Test System Based on Modified Elman Neural Network

作　　者：宋崇智[1,2] 赵又群[1] 

机构地区：[1]南京航空航天大学,南京210016 [2]安徽工业大学,马鞍山243002

出　　处：《中国机械工程》2016年第1期1-6,共6页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(11072106)

摘　　要：提出了一种能满足多层网络、多阶系统的改进型Elman网络,建立了基于改进Elman神经网络的PAC控制器。对六自由度悬架试验平台系统进行了控制研究,分析了悬架参数对轮荷利用率和相位角的影响。整车实验证明:参数匹配的悬架可以有效减小车身振动,降低悬架动挠度和轮胎动载荷。An improved Elman neural network controller was established, which could meet multi network and multi order system. A programmable automation controller（PAC） controller was established based on modified Elman neural network, and was used to the 6-DOF suspension system test platform control; the effects of suspension parameters on wheel load utilization ratio and phase angle were analyzed. The results show that the matching of the suspension parameters can improve body vi- bration, reduce the suspension dynamic deflection and tire dynamic load.

关 键 词：悬架 六自由度 试验系统 神经网络 

分 类 号：U467.5[机械工程—车辆工程]