摘要

A permanent magnet linear synchronous motor (PMLSM) is sensitive to various disturbances such as parameter variations, external load disturbances, end effect and so on. Therefore, its model is a nonlinearity, multiple-input multiple-output, strong-coupling and high order system, and its precision and speed tracking rapidity are affected by the coupling between variables. A novel linearization control scheme named neural network inverse (NNI) is proposed to overcome this problem. First, the dynamic mathematical model of the PMLSM drive system is given and its reversibility is proved. Second, a static NN is employed to identify the inverse model of the PMLSM drive system, and the static NN and two integrators are together constitute the NNI. Moreover, a pseudo-linear system can be obtained by combining the NNI with the PMLSM drive system, and then the original system is equivalent to a 2-order linear integral subsystem. Third, a proportion differentiation (PD) controller is designed to achieve the close-loop control of the PMLSM drive system. Simulation results confirm that the proposed NNI control scheme has good performance in speed and position tracking, and strong robustness against parameters variations and external load disturbances.