Research On The Control System Of A 3-DOF Precision Motion Stage Driven By Piezoelectric Actuators

As people’s requirements for precision machinery increase day by day,micropositioning technology becomes more and more important.One of the mainstream research directions in micro-positioning technology is the development of micropositioning stage.Among them,the micro-positioning stage system driven by piezoelectric actuators has high precision,high resolution,high reliability,fast response,easy processing and control.And other advantages,it is usually used in precision equipment,instruments and precision operations,and has important value in scientific research and production applications.In order to solve the positioning problem of a well-designed 3-DOF precision motion stage and make it perform well,a control method for this device is proposed,which combines feed-forward control based on neural network and PID-based feedback control.Since the device realizes three-dimensional motion,the displacement trajectories in three directions are converted into the motion trajectories of each piezoelectric actuator through the kinematic model of the device,and then the motion of each piezoelectric actuator is controlled separately.Firstly,the kinematics model and inverse kinematics model of the device were established and parameters were calculated,and the correctness of the kinematics model was verified by using finite element analysis software.Then,the driving model of the device was established and the parameters were calculated,and the output of the piezoelectric actuator was connected with the input of the device to prepare for the design of the subsequent control system model.Secondly,a hysteresis nonlinear compensation method for piezoelectric actuator based on BP neural network is proposed to further design the control system model.Firstly,the input and output of the neural network are determined through the PrandtlIshlinskii model,and then the data set for training the neural network is collected and the neural network is trained,and the optimal number of neurons in the hidden layer is obtained.Finally,the control effect of the piezoelectric actuator was tested by using the trained neural network.The experimental results show that using this method to control the piezoelectric actuator can make its output displacement basically the same as the theoretical displacement,which verifies the effectiveness of the neural network to compensate the hysteresis nonlinearity of the piezoelectric actuator.Finally,the control system of the 3-DOF precision motion stage is built,and combined with the previous work,Simulink is used to design the control system model for DSPACE.After the preparation is completed,the PID controller parameters are tuned by using the critical oscillation method combined with the influence weight of each PID controller on the deformation of piezoelectric actuators.The test results of the control system show that within the scope of application of the control system,the motion trajectory output by the device is basically consistent with the set trajectory,which verifies the effectiveness of the control system.