Research On Servo Speed Control System Of Permanent Magnet Synchronous Motor

Permanent magnet synchronous motor(PMSM)has the advantages of high efficiency,small torque ripple,good dynamic response and wide speed range.So it is widely used in national defense,industry,transportation,life and other areas.PMSM servo speed control system usually uses vector control(VC)and direct torque control(DTC).The torque response of VC system is slow.VC system is also affected by the rotor parameter changes.DTC system is simple,but it has a large torque ripple.The traditional PI controller parameters are less robust in the servo speed control system of PMSM.It can't meet the requirement of system steady-state precision.At present,the nonlinear theoretical research of model predictive control(MPC)is more active.In this paper,the neural network(NN)and support vector machine(SVM)are combined with MPC respectively.The MPC structure and algorithm of PMSM servo speed control system are designed on the basis of VC.It is expected to give a full play to the advantage of MPC to improve the overall performance of PMSM servo speed control system.The main contents of this paper are as follows.(1)The robustness of PI parameters is poor in the PMSM speed control system.This paper takes method to improve the problem.Firstly,the parameters of the PI controller are not optimal by the engineering design method.So they are optimized by the seeker optimization algorithm(SOA).Secondly,using the neural network PI control.When the operating state of the system changes,the parameters of the PI controller are automatically adjusted by the neural network,and the parameters of the PI control are optimized by the self-learning and modifying the parameters of the neural network.(2)Considering the nonlinearity and uncertainty of PMSM servo speed control system,the prediction model of PMSM is established by RBF(Radial Basis Function)neural network with the basic principle of MPC.The speed control system of PMSM is designed with RBF neural network predictive control on the basis of vector control.According to the method of multi-step predictive control,the prediction model of PMSM is established,and the optimal control quantity is optimized by the gradient descent method.The robustness and dynamic performance of the servo speed control system are improved.(3)The prediction model of PMSM is established by least square support vector machine(LS-SVM).The speed control system of PMSM with LS-SVM predictive control is designed on the basis of VC.This paper analyzes the principle and method of LS-SVM of regression prediction,and designs the nonlinear controller by Newton method,which improves the response speed and steady-state accuracy of PMSM servo speed control system.(4)The dynamic model of the mechanical transmission part is established by analyzing the structure of the servo feed system of the CNC machine tool.The PMSM speed control system with PI control and RBF neural network predictive control is applied to the CK6136 CNC lathe feed servo system,which realizes the precise control of the table position.