Neural Network-based Model Optimization And Control Of Piezoelectric Stick-Slip Actuators

As an important research direction in nanoscience,precision positioning devices and related technologies with micro-and nano-level positioning accuracy have been widely used in the fields of scanning electron microscopy,biomedical science and microelectromechanical systems.Piezoelectric stick-slip actuators are often used in the field of nanopositioning due to their compact structure,good stability and high repeatable positioning accuracy.However,in practical applications,the piezoelectric ceramic material characteristics and operating principles lead to complex non-linear characteristics between the input signal and output displacement of the piezoelectric stick-slip actuator,with problems such as steady-state positioning errors and poor anti-interference capability,which affect the control accuracy of the positioning platform.To solve the above problems,this paper takes the piezoelectric stick-slip actuator platform as the research object and designs the closed-loop positioning control of the nano-precision positioning platform to achieve the positioning control target of high precision and long stroke.Specific research includes the following aspects:Study of the piezoelectric stick-slip drive principle and non-linear characteristics.By analysing the working principle of the stick-slip drive,the non-linear characteristics of the piezoelectric stick-slip drive are explored in depth to provide a theoretical basis for the establishment of an accurate mathematical model.The piezoelectric stick-slip drive positioning system is built based on MATLAB and Lab VIEW,and the various non-linear influencing factors of the system are analysed as the necessary basis for subsequent model building and controller design.Establish a piezoelectric stick-slip actuator dynamics model.In order to accurately describe the motion of the piezoelectric stick-slip positioning table,various hysteresis models and friction models were compared and analysed,and a suitable model was selected according to the actual situation of the positioning table to establish a comprehensive dynamics model of the piezoelectric stick-slip drive.MATLAB software is used to simulate and analyse the dynamics model to provide guidelines for the design of the closed-loop controller of the piezoelectric stick-slip actuator and to verify the accuracy of the established comprehensive dynamics model.Design of a closed-loop positioning control experimental scheme.A combined macro and micro controller is designed for the motion characteristics of the positioning stage,with a step mode control in the macro motion phase and a scan mode control in the micro motion phase.To achieve high accuracy positioning control of the target in the micro-motion stage,a dynamic sliding mode controller based on RBF neural network is proposed,and the stability of the piezoelectric stick-slip driven positioning system is analysed using Lyapunov stability theory.To solve the time delay problem in the whole control system,the Lambert W function is proposed to optimise the controller parameters of the closed-loop control system.Finally,the feasibility and effectiveness of the macro micro-positioning control method is verified through an experimental platform.