自适应神经元实现的直线永磁同步电机伺服系统的预见前馈补偿

介绍了高精度微进给直线永磁同步电机伺服系统的 IP位置控制 ,提出了对该系统的最优预见前馈补偿 ,以提高系统的跟踪性能。为了能适应系统参数的变化 ,采用自适应神经元来实现预见前馈补偿。自适应神经元的输入向量为预见步数 ,权值为预见前馈系数 ,权值具有明确的物理意义 ,而且权的初始值不是随机数 ,而是系统在额定情况下算出的预见前馈系数 ,从而加快了神经元的调整速度 ,提高了系统的跟踪能力。仿真实验结果证明了所提出方案的有效性

After presenting briefly IP position control for a high-precision micro-feed linear permanent magnet synchronous motor(LPMSM) servo system in this paper, we present a method of optimal preview feedforward compensation for this system to improve the tracking performance. To adapt the parameter variability of this system, we adopt Adaline to implement preview feedforward compensation. The input vector of Adaline is the preview steps, and its weight value is the preview feedforward coefficient. It is obvious that the weights have explicit physical meaning, and that the initial values of the weights are not random, but are the preview feedforward coefficients worked out under the rated conditions, so the convergence speed of Adaline is very fast. To demonstrate the significance of the proposed method, simulations of this application are carried out.

国家自然科学基金资助项目(59875061)