神经网络在直线电机进给系统建模中的应用

获取直线电机伺服驱动系统的非线性模型，对这种先进的进给系统的广泛应用有着重要的意义。在对非线性模型建立方法分析的基础上，应用Levenberg-Marquardt优化策略，对前馈BP神经网络建模过程进行了推导。并利用试验获取的^数据，以直接辨识的方式，运用推荐的算法建立了某类型直线电机运动平台的非线性闭环系统模型。在模型的基础上仿真了运动系统的阶跃响应。由响应曲线表明，所获得的模型与实际工况中的系统很相符。

Modeling of Linear Motor Feeding System by Neural Network

Getting the nonlinear model of the linear motor servo drive system,is of importance for the abroad application of the advanced feed system.On the analysis of the estimates for nonlinear model,the set-up process of back propagation(BP) neural network is deduced with the exploitations of the optimal strategy Levenberg-Marquardt algorithm.And by the direct way,the nonlinear closed-loop model of a certain linear motor motion table is founded with experimental data.The step response of the motion system is simulated with the model above.The response curve indicates that the achieved model greatly agrees with the system in the real operating mode.