Research On Sensorless Control For PMSM System Based On Extended State Observer

At present, permanent magnet synchronous motor(PMSM), which possesses the advantages of simple structure, high power density, high efficiency, smooth operation and etc., has become the first choice in the small or medium-sized power field of high-precision servo drives. Vector control is the most widely used in industrial applications, it can meet the precise control of speed and current,. In order to ensure high performance of the PMSM control system, need to obtain accurate rotor position and velocity information at any time, generally install mechanical sensor on the rotor shaft for the direct measurement of motor speed and position to achieve high-performance closed-loop control. The installation of high precision position sensor not only reduces the system's reliability and anti-interference ability, but also increases the volume weight and cost of the control system. Therefore, the research of sensorless control technology has become a hot and key technology. In this paper, the application and implementation of sensorless control based on observer for a PMSM control system have been studied.Firstly, based on the design theory of observer, this paper introduces the non-smooth feedback control compared with linear feedback control in accuracy and dynamic performance, designed the first-order linear state observer, sliding mode observer and improved sliding mode observer, the experiment result compared and analyzed. Then extended state observer(ESO) was proposed to overcome slow response of first order dynamic observer and chattering problem of sliding mode observer. In addition, this paper analyzed quadrature phase-locked loop speed signal extraction method and putted forward the improvement measures using state differential structure to replace the traditional PI structure. Through theoretical analysis and experiments verified that the improved extraction method could enhance the accuracy of observation and noise anti-jamming performance.Secondly, this paper focused on sensorless starting process and the smooth transition process. I/F single current closed-loop control was applied to realize the PMSM sensorless stating up strategy by comparing the shortcomings of the conventional V/F method. Finally, this paper analyzed and studied the dynamic process of the sensorless control from low speed area to the high speed area. Two methods based on the acceleration of virtual coordinate system and the principle of self-balancing of the motor were used to realize the smooth transition of the transition process. The advantages and disadvantages of the two methods are compared by experiment.