汽车磁流变减振器神经网络模型研究  

Study of neural network modelling of automotive magnetorheological fluid shock absorber

作　　者：廖昌荣[1] 余淼[1] 杨建春[1] 陈伟民[1] 黄尚廉[1] 

机构地区：[1]重庆大学光电工程学院,重庆400044

出　　处：《中国公路学报》2003年第4期94-97,共4页China Journal of Highway and Transport

基　　金：国家自然科学基金重点项目(50135030);重庆市院士基金项目(027754)

摘　　要：提出了用3个输入单元、1个输出单元和7个隐含单元的神经网络来摸拟汽车磁流变减振器的动力学特性,采用实验室测试得到的数据和基于高斯—牛顿法的改进算法对神经网络的连接权值进行优化。神经网络预测结果与试验测试结果比较表明:提出的神经网络模型能够比较准确地预测磁流变减振器的动力学特性。Authors present a multilayer perceptron neural network with 3 input neurons, 1 output neuron and 7 neurons in the hidden layer to simulate the dynamic behavior of automotive magnetorheological fluid shock absorber. Optimization training for weighting coefficients of the neural network model has been done by GAUSSNEWTON method improved in this paper using data generated from test in laboratory. In comparison with experimental results of magnetorheological fluid shock absorber, the neural network models are reasonably accurate over a wide range of operating conditions.

关 键 词：汽车工程 减振器 神经网络 磁流变体 

分 类 号：U463[机械工程—车辆工程]