基于神经网络的磁流变减振器模型参数辨识  

Model Parameter Identification of Magnetorheological Shock Absorber Based on Neural Network

作　　者：张宇涵 陈双[1] 任洪卓 Zhang Yuhan;Chen Shuang;Ren Hongzhuo(Liaoning University of Technology,Jinzhou 121000)

机构地区：[1]辽宁工业大学,锦州121000

出　　处：《汽车技术》2022年第9期50-55,共6页Automobile Technology

基　　金：国家自然科学基金项目(51605213);辽宁省教育厅科学研究项目(JJL202015402)。

摘　　要：为了建立精确的磁流变减振器模型,针对单筒单出杆式磁流变减振器开展力学性能特性试验,采集其在不同性质激励下的阻尼力、位移、速度数据,得到磁流变减振器的工作特性;应用具有7个神经元输入层的BP神经网络搭建磁流变减振器的正向模型和逆向模型,并通过试验数据对模型测试集进行仿真验证。结果表明,所采用的7个神经元输入层BP神经网络建立的模型可以精确预测磁流变减振器的输出阻尼力和控制电流,具有较好的可靠性。In order to establish an accurate model of magnetorheological shock absorber,the mechanical performance characteristic test is conducted for single-cylinder single-lever magnetorheological shock absorber,and the damping force,displacement and velocity data are collected under different excitations to get the working characteristics of magnetorheological shock absorber.BP neural network with 7 neuron input layers is used to build the forward model and inverse model of magnetorheological damper,and the model test set is simulated and verified by test data.The simulation results show that the model established by BP neural network with 7 neuron input layers can accurately predict the output damping force and control current of magnetorheological damper,it has good reliability.

关 键 词：磁流变减振器 BP神经网络 正向模型 逆向模型 

分 类 号：U461.4[机械工程—车辆工程]