基于经验模态分解算法的永磁直线同步电机迭代学习控制

摘要

在永磁直线同步电机驱动伺服系统的迭代学习控制(ILC)过程中,针对由于每次运行时跟踪误差的累积,导致系统出现收敛速度降低甚至发散的现象,提出一种基于经验模态分解(EMD)算法的迭代学习控制方法.首先设计闭环ILC控制器,然后利用EMD算法分解ILC过程中的跟踪误差,筛选并消除其中发散的分量,保证ILC的收敛性,提高ILC的收敛速度.仿真和实验结果表明,与传统ILC相比,所提出的控制方法能够使系统的跟踪效果更好,且保证了伺服系统的输出轨迹在较少的迭代次数下快速精确地收敛到期望轨迹.%In the process of iterative learning control (ILC) of the servo system driven by permanent magnet linear synchronous motor (PMLSM), the accumulation of tracking error would lower convergence rate or even system divergence. In this paper, a closed-loop ILC controller was designed, and then a method that combined empirical mode decomposition (EMD) algorithm with ILC was adopted to decompose the tracking error in ILC. Then the divergent components were screened and eliminated. Accordingly, the convergence of ILC was guaranteed, and the convergence rate was improved. The simulation and experimental results show that, compared with the traditional ILC, the presented method can produce better tracking results, and guarantee that the output trajectory track the anticipant trajectory quickly and precisely with fewer iterations.