Research Of Sensorless Control Of PMSM Based On Kalman Filter

Permanent Magnet Synchronous Motor(PMSM) has been widely used in themotor drive fields because of its superior operating performance, small size, etc. Thevector control method is often used since it is convenient and reliable.Using vectorcontrol needs to know the information of the rotor.The traditional method is using aspecial mechanical sensor to obtain information.However,the mechanical sensor iseasily affected by external factors.At the same time,using it increases cost of themotor control and decreases motor control accuracy and limites the application of themotor.Because of that, how to get the rotor information without the aid of amechanical sensor has become a hot research point. The control method which do notuse mechanical sensors is called sensorless control method.In this paper, thesensorless control method is researched deeply.Firstly,this paper discusses the background and development status of thesensorless control.Several major sensorless control methods are introduced.Secondly,mathematical models of PMSM are described detailed.Three commonmotor mathematical models in different coordinate system are described andanalyzed which are their advantages and disadvantages,which pave the way of themotor sensorless control.Then,depth study of control algorithms for the sensorless control applying todifferent speed range is done.First，Kalman filter which applying to the high-speedstage is studied.A new motor model is established to avoid the disadvantage of theextended Kalman filter which needs large computation.The reduced-order linearKalman filter algorithm is given to estimate the information of the rotor.Using thisalgorithm can not only reduce the calculation amount, but also ensure the estimationaccuracy.Then rotating high-frequency voltage injection method and pulsatinghigh-frequency voltage injection method are mainly studied., based on Finding theshared character of these two methods,stator current unified model is established.Acontrol algorithm based on the model which combine high-frequency signal injectionmethod and the Kalman filter is applied.Finally, sensorless control of PMSM based on the foregoing analysis issimulated.The effectiveness of proceeding algrithm is verified.A suitable switchingscheme for high-speed and low-speed algorithm is given,and is proved to be effective.