Research On Sensorless Control Strategy Of Permanent Magnet Synchronous Motor In Full Speed Range

Permanent magnet synchronous motor(PMSM)has the advantages of high efficiency and small volume,etc.And it has been widely used in railway industry and industrial automation.In order to implement close-loop control of PMSM,position sensor is needed to detect rotor position.Additional position sensor not only increases costs but also reduces system reliability.Therefore,PMSM sensorless control has become a hot research topic in this field.Aiming at the problem of initial rotor position detection of PMSM,this paper studied the estimation methods applied to zero speed and non-zero speed.The successive approximation method is often adopted at zero speed,but its detection accuracy is susceptible to inverter nonlinear factors.In order to solve this problem,two voltage vectors of different values is injected in the same direction,and the estimated angle is carried out using the difference of two response current.When the motor spins,this paper analyzes the factors leads to estimation error of short circuit method based initial position algorithm.The method of selecting experimental parameters(short circuit time and short circuit interval)is proposed,which can limit the influence of sampling error and parameter uncertainty to a smaller range at different speed and improve the estimation accuracy.Three algorithms based on high frequency voltage injection are introduced,including high frequency rotating voltage injection method,high frequency pulsating voltage injection method and high frequency square voltage injection method.The injection method,processing method and implement method are analyzed in detail.And also the elements which influence the estimation precision is studied.This kind of methods requires digital filter to separate signals of different frequency.This paper analyzes the effect of the phase delay of the filter on the estimation results,and proposed the method to compensate phase delay.The accuracy of rotor position estimation is greatly improved.The phase-locked loop of the traditional algorithm is replaced by the luenberger observer.The problem of static estimation error of rotor position when the motor speed changes as slope.The algorithm and improvement strategy mentioned in this paper are simulated and verified by Matlab/Simulink software and 3kw traction platform.Experimental results demonstrate the validity of the theoretical analysis and the effectiveness of the algorithm in this paper.