Intelligent Second-order Sliding Mode Control For Permanent Magnet Linear Synchronous Motor Servo System

The rapid growth of high-performance permanent magnet materials and the constant improvement of power electronics technology make rapid progress in permanent magnet linear synchronous motor(PMLSM).Because PMLSM possesses the merits of no mechanical friction,simple mechanical structure,large thrust and small volume,it has fitness for high speed and high precision servo applications.Nevertheless,due to the absence of gear set,ball screw and other indirect mechanical transmission links,the PMLSM servo system is more easily to be affected by the time-varying uncertainties such as external disturbances,parameter variations and nonlinear frictions.In order to overcome the adverse effects of uncertainties on the system,an intelligent second-order sliding mode strategy is put forward by combining the sliding mode control method with strong robustness and the neural network control method with strong learning ability.First of all,the fundamental structure and the operating principle are elaborated in detail,and the mathematical model of PMLSM containing uncertainties is established.Then the PMLSM vector control system is established,and the cause of uncertainty in the system is analyzed.Then,to address the problem that the system is easy to be affected by the external disturbances,parameter variations,nonlinear friction and other uncertain factors,a second-order sliding mode control(SOSMC)system is adopted by means of PID type second-order sliding surface,which ensures the finite-time convergence of output tracking errors and enhances the control performance.Due to the chattering problem in sliding mode control,the quasi sliding mode control method is adopted,and the saturation function boundary layer control method and sinusoidal saturation function boundary layer control method are employed respectively to weaken the chattering.The simulation results show that the SOSMC methods are able to significantly weaken the chattering,effectively improve the position tracking accuracy,and enhance the robustness of the system.Finally,to address the problem that it has trouble to approximate the external disturbances,parameter variations and nonlinear friction in the system,which makes it difficult to select the parameters of the system,neural network is adopted to estimate the uncertainty,so as to enhance the robustness of the system.On the foundation of the radial basis function neural network(RBFNN),intelligent SOSMC methods combining recurrent radial basis function neural network and dynamic recurrent radial basis function neural network with SOSMC are developed respectively to bring about the optimal performance of the control system.The simulation results demonstrate that,compared with SOSMC,intelligent SOSMC possesses fast and precise position tracking performance,strong anti-interference ability and superior control effect for PMLSM servo system.