Self-Learning Compensation of Hysteretic and Creep Nonlinearities in Piezoelectric Actuators

摘要

Solid-state actuators based on active materials allow high operating frequencies with nearly unlimited displacement resolution. This predestines them, for example, for application in highly precise positioning systems. However, the nonlinear behaviour in such systems is mainly attributed to actuator transfer characteristics as a result of driving with high amplitudes. In this paper a novel self-learning compensation method based on the so-called modified Prandtl-Ishlinskii approach is presented, which allows extensive compensation of the complex solid-state actuator hysteretic and creep nonlinearities during operation. Finally, it is shown in an example involving a two-axis piezoelectric parallel-kinematic positioning system that this compensation substantially decreases the deviations of the actual output displacements from the desired displacements of the controlled system.