High-precision Motion Control Of Piezoelectric Actuator System Based On Active Disturbance Rejection And Current-cycle Iterative Learning

As a typical nonlinear characteristic,hysteresis nonlinearity which widely exists in the modern industrial field,has attracted the attention of many scholars,and has become one of the hot research directions.In order to meet the nano-level precision requirements of micropositioning,micro-operation and micro-vibration control in aerospace,national defense technology,precision microelectronics and other industrial production,it is urgent to study the modeling and control of dynamic hysteresis nonlinear system.However,in practical engineering applications,the hysteresis nonlinear system often exhibits a compelex multifactor coupling dynamic characteristic,which not only will limit the control accuracy of the system,but also may cause system instability or even lead to system oscillation,in the relationship between input and output.The traditional nonlinear theories and control methods have been unable to cope with and solve this complex problem well.Therefore,a piezoelectric actuator(PEA)composed of smart materials and a six-degree-of-freedom piezoelectric Stewart platform with cubic structure are used as control objects in this study.The modelling and high-precision motion control method of a rate-dependent hysteresis nonlinear system are studied.The effectiveness,superiority and engineering applicability of the proposed methods are verified by numerical simulations in MATLAB-Simulink and comparative experiments on d SPACE platform.The main work contents of this paper involve the following items:(1)Based on the electrical/mechanical actuation mechanism of the PEA,and taking the rate-dependent hysteresis nonlinearity characteristics of the system as external disturbances,a second-order disturbance-based(SODB)comprehensive dynamic model of the PEA,which can get rid of the dependence on the specific hysteresis model of the system and avoid the complex and tedious hysteresis modeling process,is proposed.The structure of the model is simple which is convenient for the design of the controller;(2)Based on the SODB model,an active disturbance rejection control(ADRC)method for dynamic hysteresis compensation of the PEA is proposed.A linear extended state observer(LESO)is used to estimate the state and disturbance of the system on-line in real time,and control laws are designed to eliminate and control the hysteresis of system.It is verified that the proposed method can effectively compensate for the rate-dependent hysteresis nonlinearity characteristics of the system through the tracking control experiments of PEA under multifrequency and multi-type reference signals,and comparing with pure PI control;(3)Based on the ADRC for hysteresis compensation of the system,a current-cycle iterative learning control(CILC)method for high-precision motion of the PEA is proposed,and the learning convergence conditions of the system error are analyzed and derived in the iterative domain.The tracking control experiments are carried out under various types of reference signal input in the frequency range of 1-200 Hz,and compared with nonlinear proportional-integral(NPI)control and adaptive sliding mode(ASM)control.The experimental results show the effectiveness and superiority of the proposed method,which significantly improves the control performance and greatly reduces the tracking error of the system.The smallest maximum absolute error(MAE)is 16.3nm,which can meet the needs of the actual engineering;(4)Based on the high precision motion control of the single-degree-of-freedom PEA,an ADRC method for high-precision motion of the six-degree-of-freedom cubic structure piezoelectric Stewart platform is proposed.The geometric and dynamic analysis of the Stewart platform are carried out,and the inverse kinematics and forward kinematics solutions of the platform were designed.The model of the PEAs which are used to drive the legs is added in the process of establishing the comprehensive dynamic model of the whole system,and the effect of the inherent rate-dependent hysteresis nonlinearity characteristic of the PEA is introduced.The effectiveness of the proposed method is proved by the MATLAB-Simulink simulation results.