Iterative Learning Synchronous Control Of Linear Motor Drive H-Type Platform

At present,the advantages and disadvantages of CNC machine tools mainly depend on accuracy and speed.The development direction of automation equipment tends to be high-speed and high-precision in the future.Permanent magnet linear synchronous motors（PMLSM）are widely used in precision device processing and other applications because of their high repeatability and reliability.Due to PMLSM eliminates mechanical transmission and is directly connected with load,it is sensitive to external disturbances.The direct-drive H-type platform is mainly composed of three PMLSM,including one permanent magnet linear synchronous motor in the X direction and two parallel permanent magnet linear motors in the Y direction.However,in H-platform system,uncertain disturbances such as time-varying parameters on one axis,external disturbances or coupling nonlinearity on two axes will affect the performance of the system and reduce the tracking accuracy.The main purpose of this thesis is to reduce the synchronization error between Y1 and Y2 axes of H-type platform by designing a learning model predictive controller（L-MPC）on one axis and an iterative learning controller（ILC）on two axes.Firstly,the development status of H-type platform at home and abroad is summarized,and the main control methods of synchronous control of PMLSM and direct-drive H-type platform and the factors affecting the positioning accuracy of the platform are discussed,and the mathematical models of PMLSM and H-type platform are deduced and established.Secondly,aiming at the single-axis position error of linear motor,the model predictive control（MPC）is applied to the single-axis position loop.Through the predictive model,rolling optimization is carried out with the best performance index,and then feedback correction is carried out to reduce the position error.L-MPC is proposed to further improve the tracking performance of reference signals.,which is suitable for ILC of repetitive motion to update the input of MPC.The simulation models of MPC and L-MPC are built in Matlab/Simulink respectively.The simulation results show that L-MPC can effectively reduce the position error of single axis on the basis of MPC.Finally,in order to reduce the synchronization error between two axes,based on the single-axis controller,ILC that is suitable to repetitive motion is designed between two axes to reduce the synchronization error of double linear motors on Y axis caused by uncertain factors.At the same time,in order to further improve the control accuracy and calculation speed,and reduce the number of iterative learning,fractional order control is introduced to form fractional order iterative learning control（F-ILC）,which is developed on the basis of integer order and has better adjustment effect.Variable gain fractional-order iterative learning control（VGF-ILC）is proposed to be applied to biaxial system,and the convergence speed of the system can be improved by changing the learning gain.The simulation models of ILC,F-ILC and VGF-ILC are built in Matlab/Simulink.The simulation results show that F-ILC and VGF-ILC greatly improve the control accuracy and convergence speed between two axes of the system.