Research On Online Parameter Identification Of PMSM Based On Two Steady States

Permanent magnet synchronous motors(PMSMs)are attracting more and more attention in rail transit systems because of their small size,high power density,and excellent control performance.It is important to obtain accurate motor parameters(?_f,R_s,L_d,L_q)for high-performance control.However,parameters are affected by temperature variation and magnetic saturation,which will result in poor control performance,especially for the accuracy of actual output torque of PMSM.Therefore,it is necessary to identify the parameters online.This paper focuses on the method for simultaneously identifying all four parameters of salient-pole PMSM.The main contents are as follows:The mathematical model of PMSM in dq-axis coordinate system is derived and the influence of temperature variation and magnetic saturation on motor parameters is analyzed.Then,the vector control system is introduced based on the mathematical model.Maximum torque per ampere(MTPA)control,which is the most common method to maximize the efficiency of the system,is introduced for the subsequent analysis of multiparameter identification method.In order to solve the rank-deficient problem of online multiparameter identification,the conditions for simultaneously identifying four parameters are derived.This paper proposes a method where a temporary steady state is constructed to meet the full-rank conditions.The theoretical error,which is caused by the variation of parameters,is theoretically derived and analyzed with different output torques.Meanwhile,the best shift direction of the temporary operating point is designed based on the analysis of the theoretical error.In order to reduce the influence of non-ideal factors such as initial position angle error,control delay and dead zone,the errors caused by the non-ideal factors are derived to show the magnitude of the effect.After that,several methods have been proposed to correct or compensate these non-ideal factors,which will improve the accuracy of the identification method.At the end of this paper,the proposed method is validated in the simulation and experimental platforms of salient-pole PMSM.The results show that all four parameters can be simultaneously identified based on the proposed method.It has good performance in multiparameter identification.Meanwhile,these non-ideal factors,which have great influence on the proposed method,can be corrected or compensated for better identification performance.