Research On Modulated Model-free Predictive Control For PMSM Drive System

Permanent magnet synchronous motor(PMSM)drive system is the main power source of new energy vehicles,whose control quality directly affects the performance of the new energy vehicles.With the advantages of simple structure,small size,high efficiency and high power density,PMSM has been widely used in electric drive system of EVs.Therefore,this paper focuses on the key technologies in PMSM drive system,such as the improvement of steady-state control performance and the reduction of switching losses.The research of modulated model-free predictive control,modulated model-free predictive control with minimum switching losses,and modulated model-free predictive control based on DSVM is presented,which has theoretical research value and engineering application value for improving the current control performance of PMSM drive system in EVs.Firstly,this paper introduces the principle of modulation,and the current ultra-local model of PMSM drive system is established.Then the modulated model-free predictive control of PMSM driving system is proposed.By analyzing the vector properties of the current error predicted by the eight basic voltage vectors of the inverter,the linear combination method is used to make the stator current error equal to zero at the next control period,and realize no-error stator current tracking.Aiming at the problem of dynamic control performance degradation under the condition of over-modulation operation,the over-modulation scheme based on the principle that minimum predictive current error amplitude is studied.Then,from the angle of current vector modulation,a modulated model-free predictive control with minimum switching losses for PMSM drive system is proposed,which makes the current vector at the next control period is equal to the reference current vector by reasonably allocating the duty cycles of the predictive current vectors of the inverter to obtain the optimal voltage vector of inverter.Then,a rotating coordinate frame is used to adjust the optimal voltage vector,reducing the complexity of over-modulation operation.And then,by calculating the switching losses of the vector sequences with different zero voltage vectors,and taking the switching losses as the second-level cost function to select the optimal voltage vector sequence,the switching losses of the inverter can be reduced.Finally,combining the DSVM technology with the modulated model-free predictive control,two modulated model-free predictive control based on DSVM are proposed.According to the duty cycle of the basic voltage vector obtained by the modulated model-free predictive control,the characteristics,the advantages and the disadvantages of two different online optimal voltage vector determination methods are analyzed.Then through the idea of hierarchical control,the sequence of voltage vector is optimized.On the basis of theoretical analysis,through the simulation and experimental study of the SMPMSM driving system,the effectiveness of the proposed modulated model-free predictive control,modulated model-free predictive control with minimum switching losses and modulated model-free predictive control based on DSVM are verified.