Research On Single Control-loop Model Predictive Algorithm For Permanent Magnet Synchronous Motor Servo System

Permanent magnet synchronous motor(PMSM)has become the main actuator of AC servo system due to its advantages of wide speed range,high efficiency and easy maintenance.However,owing to the characteristics of multi-variables,strong coupling and non-linearity,and the existence of parameter uncertainties,the load disturbance and friction in the PMSM servo system,it is difficult for the traditional linear control methods to guarantee a high control performance.In recent years,with the development of control theory and microprocessor technology,many advanced control methods have been proposed and applied to motion control field.As a kind of optimal algorithm,model predictive control(MPC)can make full use of model information and deal with constraints effectively.Besides,compared with cascade control methods,single control-loop model prediction algorithm replaces the speed loop and current loop with a control loop,which has the advantages of simple structure,few parameters to be adjusted and fast dynamic response.Therefore,in order to improve the control performance,this paper will study the single control-loop model prediction algorithm in the PMSM servo system.Based on the vector control strategy,a single-loop MPC method is designed according to the second-order equation between voltage and speed of PMSM.Furtherly,in order to improve the speed loop bandwidth,a composite control method based on single-loop MPC and reference signal feedforward is proposed.Simulation and experimental results show that the single-loop MPC method has better dynamic and steady performances than the cascade PI control.Besides,the proposed composite method can enhance the speed bandwidth.For the problem of PMSM parameter variation and load disturbance in the actual working condition,a single-loop MPC based on disturbance observer(DOB)is studied.Firstly,a second-order linear DOB is designed to estimate the lumped disturbances consisting of model parameter uncertainties and external load disturbances.Then,the disturbance estimation is introduced into the prediction model to obtain accurate prediction output,and a single-loop MPC controller is established based on this model.Simulation results verify that the robustness and disturbance rejection ability of this proposed method are strong.Considering that friction in the actual system can cause the degradation of low speed tracking performance of PMSM,so a single-loop MPC method based on friction compensation is proposed.Firstly,the friction model is identified by the least square method,and then a DOB is used to estimate the lumped interference composed of friction compensation deviation and load disturbance.At last,the friction torque and disturbance estimation are introduced to establish a more accurate prediction model,and a MPC law is designed on this basis.It is shown by simulation and experimental results that a better speed tracking performance in low speed and round-trip motion can be obtained.