Research On Multi-Step Model Predictive Control Optimization Of Virtual Voltage Vectors For Permanent Magnet Synchronous Motor

Permanent magnet synchronous motor has many advantages such as small size,high efficiency,high torque-to-inertia ratio and strong reliability.Currently,permanent magnet synchronous motor is widely used in aerospace,new energy Automobile,high-precision servo system,rail transportation traction and other fields.Field orient control and direct torque control are two traditional control strategy for permanent magnet synchronous motor drives.However,these two methods are difficult to meet the requirements in some occasions with high real-time and high performance requirements because these two methods both rely on off-line PI controllers.With the rapid development of power electronic technology and control theory in recent decades,model predictive control theory has been successfully combined with high-frequency power electronic devices and widely applied in the field of AC motor drives.In this paper,the principle of model predictive control is firstly analyzed,and then the finite set model predictive control and continuous set model predictive control are respectively studied.Compared with the continuous control set model predictive control,the finite control set model predictive control has lower control performance,but its structure is simple and the calculation amount is small.Based on the mathematical model of permanent magnet synchronous motor,the implementation process of finite control set model predictive torque control is studied from the aspects of cost function construction and prediction model.However,the traditional finite control set model predictive torque control for permanent magnet synchronous motor drives has problems such as high torque ripples and high current harmonic components.In order to solve these problems,the model predictive control based on virtual voltage vector is studied in this paper.By adopting virtual voltage vector model predictive control and discrete space vector modulation technology,the switching state of the optimal virtual voltage vector can be obtained,which can improve the control performance of permanent magnet synchronous motor and reduce the calculation amount of model predictive control.Thirdly,in order to further improve the performance of permanent magnet synchronous motor system,this paper proposes two improved multi-step optimization control strategies:fixedradius multi-step optimization model predictive torque control(MS-MPTC1)and variable-radius multi-step optimization model predictive torque control(MS-MPTC2).On the one hand,MSMPTC1 reduces the computational burden by narrowing the optimal voltage vector range to one sector.In addition,MS-MPTC1 discusses the value of the virtual radius in the simulation platform based on the position of the reference vector,and the final value of the virtual radius is suitable for all working conditions of the permanent magnet synchronous motor drive system.On the other hand,MS-MPTC2 analyzes the relationship between the voltage vector amplitude and torque according to the system characteristics of the permanent magnet synchronous motor,and divides the voltage vector hexagon into three regions.MS-MPTC2 adopts different search strategies in each region to reduce the computational complexity in the predictive control process.Different from MS-MPTC1,MS-MPTC2 studies the variable virtual radius to further improve the performance of PMSM system.For the two methods,MS-MPTC1 and MS-MPTC2,the steadystate and dynamic performances of the two methods are specifically analyzed by simulation results.Finally,an experimental platform of permanent magnet synchronous motor is built.On this experimental platform,experiments on steady-state performance and dynamic performance were carried out to verify the feasibility and effectiveness of the two methods,MS-MPTC1 and MSMPTC2.The experimental results show that the two methods have excellent dynamic performance,and the computational burden has been reduced compared with other model predictive control methods,and through the quantitative analysis results of torque ripple and phase current harmonic components,the proposed two methods are effective.Steady-state performance has been significantly improved.