Research On Model Predictive Current Control Strategy For Permanent Magnet Synchronous Motor

In recent years,the applications of Permanent Magnet Synchronous Motor(PMSM)has become widespread,such as aerospace,CNC servo and electric vehicles,due to small size and high power density of the PMSM.Compared to the traditional PI control strategy,the model predictive control(MPC)strategy with the simple control principle and fast dynamic response can better meet the increasingly high control needs of many fields.This paper takes PMSM as the research object,selects the current as the optimization target,the conventional model predictive current control(MPCC)strategy,multi-vector MPCC strategy and the design and realization of the drive system for PMSM are studied respectively.First of all,the two rotor structures of PMSM are introduced from different angles.The theoretical derivation of the three coordinate systems and coordinate transformation principles between them is carried out.The mathematical model of PMSM is established under different coordinate systems.Vector control strategies are briefly introduced,including i_d=0 control methods and SVPWM techniques.Secondly,this paper presents a comparative study of the single-vector and multi-vector MPCC algorithms for PMSM.The single-vector MPCC only acts on one voltage vector in a single sampling period,so there is a large current ripple,and the steady-state performance is poor.Multi-vector MPCC(duty-cycle MPCC,three-vector MPCC)synthesizes a virtual voltage vector through multiple basic voltage vectors,and acts on the inverter as an output,which can effectively reduce current harmonics and improve system steady-state performance.The three methods are compared from three perspectives:the number of issued vectors,the selection range of voltage vectors and the calculation method of vector action time.Finally,the MATLAB/Simulink simulation model is built,and the dynamic and static performance of the three strategies are compared and analyzed.The results show that,Multi-vector MPCC has obvious advantages.Then,in order to solve the problem of poor steady-state performance of single-vector MPCC,and the problems of complex algorithm and large amount of calculation of three-vector MPCC,this paper proposes a novel three-vector MPCC algorithm considering current error.In this method,1 zero vector and 2 effective voltage vectors are applied in each sampling period,and the 2 effective voltage vectors are used to compensate the current error between the predicted value of the current generated by the action of the zero vector and the reference current:first,the number of finite control sets is expanded from 8 to 14.With 12 generalized basic vectors as the boundary,the regular hexagon is re-divided to form 12 sectors;then the current error is determined by sector,and the two optimal voltage vectors are the sectors where the current error is located.The generalized basic vector of the zone boundary;finally,the action time of the generalized basic vector is calculated according to the deadbeat control principle of the d and q-axis currents.The proposed method only needs to perform one prediction,the calculation amount is small,and it can output a voltage vector whose amplitude and direction can be adjusted.The MATLAB/Simulink simulation results prove that the method can effectively reduce the current ripple and improve the stability of the system.Finally,an experimental platform for PMSM system is built to verify whether the novel three-vector MPCC algorithm is feasible and effective.Comparing it with single-vector MPCC,experimental results show the superiority of the new approach proposed in this article.