Low Parameter Sensitivity Model Predictive Control Of Permanent Magnet Synchronous Motor

In recent years,along with the development and application of rare earth materials,the permanent magnet synchronous motor(PMSM)has been developed relatively rapidly.PMSM is currently widely used in the field of electric vehicle drive and industrial production due to its ability to achieve high-efficiency operation,low loss,and high power density.As for the control algorithm applied to PMSM,the more mature control algorithms include Field-oriented control(FOC)and Direct Torque Control(DTC).For the past few years,with the accelerated development of Digital Signal Processing(DSP),Model Predictive Control(MPC)has been gradually adopted in the field of PMSM drive control by virtue of its ease of implementation and its ability to flexibly include non-linear constraints.As a result,the MPC algorithm has become an ever-increasing research hotspot for scholars in various countries.Nevertheless,the MPC strategy also has deficiencies,for example,it has strong parameter sensitivity.When the parameters input into the prediction model do not match the parameters of the actual control object,the control effect of the MPC strategy will become unsatisfactory.Therefore,this paper will conduct research on two aspects of how to reduce the strong parameter sensitivity of MPC to model parameters.Based on the conventional single-vector model predictive current control,this paper proposes a simple single-vector model predictive current control strategy without flux linkage parameters.This control strategy mainly focuses on the inductance and flux linkage parameters in the predictive model for optimal correction.The control strategy starts by constructing a discrete integral controller to extract the inductor information based on the predicted current difference.Afterwards,the inductance parameter enters into the model are corrected by using the obtained inductance information.Furthermore,the flux linkage parameter value is calculated based on the corrected inductance value.Finally,the corrected inductance parameter value and the calculated flux linkage parameter value are substituted into the current prediction model to construct a new current prediction model.The relevant experiments designed in this paper can prove that the parameter sensitivity can be effectively reduced according to the control strategy of the constructed single-vector current prediction model without flux linkage parameters.In order to further eliminate the influence of the resistance,inductance and flux linkage parameters in the PMSM on the effect of single-vector model predictive current control,a single-vector model predictive current control strategy without motor parameters is designed in this paper.To begin with,the control strategy builds a current prediction model without any motor parameters,and only the current difference and voltage difference are contained in the new current prediction model.Moreover,this paper also analyzes the effect of the current difference and voltage difference contained in the new current prediction model on the current prediction results respectively.Next,considering that the current difference and voltage difference in the constructed current prediction model belong to different dimensions,the paper also devises relevant balance coefficients to balance their effects on the current prediction results.In the process of designing the relevant balance coefficients,this paper uses the predicted current difference to construct a discrete integral controller to achieve real-time correction of the balance coefficients.In the end,relevant physical experiments also demonstrate the effectiveness of the single vector model predictive current control strategy without motor parameters designed in this paper.