Identification and compensation control of the dynamic hysteresis model of MFC actuators

Piezoelectric macro fiber composite (MFC) has the advantages of good flexibility and strong deformability. However, the hysteresis and nonlinearity of the flexible arm driven by the MFC actuator seriously affect the positioning accuracy of the system. An improved Prandtl-Ishlinskii (PI) hysteresis model with asymmetry was proposed to solve the defects (symmetry) of the classic PI hysteresis model. The model was obtained by superimposing a series of bilateral dead zone operators with different weights and different thresholds based on the classic PI hysteresis model. The hysteresis model identification results based on the least square method show that the improved PI hysteresis model can reduce the modeling error of the MFC actuator from 16.06% to 5.58%. Besides, a discrete transfer function model of the system was established to describe the linear dynamic characteristics of the system. The discrete transfer function model and the improved PI hysteresis model were connected in series to obtain a combined model. It solves the problem that the pure hysteresis model can only describe the hysteresis characteristics under low-frequency and quasi-static conditions. Under feedforward compensation, a sinusoidal wave trajectory tracking experiment was performed on the flexible arm actuated by the MFC actuator. The measured displacement after compensation is basically consistent with the expected tracking displacement, and the tracking accuracy is over 93.62%. The experimental results demonstrate the effectiveness of the proposed improved PI hysteresis model, discrete transfer function model and compensation method. © 2022, Editorial Office of Journal of Vibration and Shock. All right reserved.