Friction Compensation Method with Iterative Learning Based Bang-Bang Compensator in PMLSM System

抄録

The nonlinear friction force is the predominant force disturbance in a permanent-magnet linear synchronous motor (PMLSM) with guide rails and it can affect position control precision, especially during motion reversals (velocity zerocrossing). To minimize the effect of nonlinear friction on position accuracy in the reciprocating motion of a PMLSM, a compensation scheme consists of an iterative learning-based bang-bang compensator (ILBBC) and a generalized proportional-integral observer (GPIO) is presented in this paper. The ILBBC is utilized to compensate for the abrupt change in Coulomb friction force in velocity zero crossing, and the GPIO is added to enhance the compensation performance for the vicious friction force. To effectively merge the ILBBC and GPIO, a three-stage compensation scheme is proposed; thus, the combination of the ILBBC and GPIO is capable of compensating for nonlinear friction in both velocity zero-crossing and non-zero regions. The design and stability of ILBBC and GPIO are theoretically analyzed. Finally, comparative experiments, which include no compensation, GPIO-based compensation, and ILBBC + GPIO compensation, are implemented under two types of sinusoidal position trajectories to demonstrate the effectiveness of the proposed compensation scheme.