基于虚拟样机和人工神经网络的选挡机构运动分析  

Motion Analysis of Gear Selection Mechanism Based on Virtual Prototyping and ANN

作　　者：张宝[1] 李健[1] 

机构地区：[1]广西科技大学机械工程学院,广西柳州545006

出　　处：《机械研究与应用》2014年第5期7-8,13,共3页Mechanical Research & Application

摘　　要：为得到汽车选挡机构工作过程中某些变量之间具体的函数关系,在ADAMS软件中建立选挡机构的运动模型并进行运动仿真分析。分别将操作杆上X方向和Y方向的作用力作为输入变量,将拉索X方向位移和摇杆转动角度作为输出变量。然后在MATLAB软件中利用人工神经网络对输入和输出变量进行多次训练。得到的函数关系图与已知参数点的线性拟合度高,为选挡机构的数值分析打下基础。To get specific functional relationship between certain variables during operation of gear election mechanism in cars, the model of gear selection mechanism was created and analyzed in the ADAMS software. The forces in X and Y direc-tions of the joystick were selected as the input variables, and the lasso displacement in X direction and rotation angle of the joystick were selected as the output variables. Then the input variables and output variables were trained in MATLAB software by ANN. The function diagram that was obtained afterward was highly fit to the linearity of the known parameters, and thus could lay the foundation for the numerical analysis of gear selection mechanism.

关 键 词：ADAMS 选挡机构 人工神经网络 模式识别 

分 类 号：U462.24[机械工程—车辆工程]