Design of Iterative Learning Controller for Permanent Magnet Linear Synchronous Motor Based on MEEMD Algorithm

Abstract

The tracking error in the iterative learning control (ILC) process of the permanent magnet linear synchronous motor (PMLSM) servo system can be decomposed into learnable and divergent components by the empirical mode decomposition (EMD) algorithm. The problem of slow convergence or even divergence due to the accumulation of tracking errors in each iteration was solved by eliminating the divergent components. However when the PMLSM system is disturbed by measuring noise, it can easily lead to mode mixing, compromising the accuracy of the decomposition. In this paper, modified ensemble empirical mode decomposition (MEEMD) algorithm is used to decompose the tracking error. The algorithm is combined with the detection of signal randomness based on permutation entropy (PE). Firstly, a suitable PE threshold is set up to remove measuring noise, and then the residual effective tracking errors signal is decomposed by EMD algorithm. Finally the divergent components are found and eliminated. From the simulation results, the divergent components are screened and eliminated more accurately, the phenomenon of mode mixing is avoided effectively, the number of iterations is reduced, the convergence speed and tracking accuracy of PMLSM servo system are improved.