Research On Position Sensorless Control For Permanent Magnet Synchronous Motor At Full Speed Range

High efficiency,high power density permanent magnet synchronous motor(PMSM)is the key power source components of advanced rail transportation equipment,aerospace equipment and other high-end equipment.The high-performance control of PMSM system requires the additional mechanical position sensors to obtain the motor’s rotor position information in real time,but the utilization of position sensor will lead to an increase in the size of the motor system,reducing reliability,etc.The position sensorless control technology abandons the mechanical position sensor and uses motor voltage and current signals to estimate the rotor position,which is feasible for applications where the position sensor cannot be installed,or works to provide redundant estimation in case of position sensor failure.Therefore,the research of PMSM system position sensorless control technology is of great theoretical and practical significance to improve the level of electrical automation of high-end equipment in China.At present,the following technical difficulties still exist in PMSM excellent performance position sensorless control,which need to be explored in depth: the initial rotor position identification is difficult to implement on the basis of rapidity and reliability under zero and low speed conditions;the accuracy of delay compensation by high frequency(HF)injection method is limited under low switching frequency;the comprehensive suppression on position harmonic error pulsation and dc bias by model method position observer;Switch the composite algorithm smoothly for full speed range operation is in urgent need,etc.This paper takes the PMSM system as the research object and conducts a profound study on the position sensorless control strategy in the full speed range.The main research work is summarized as follows.Firstly,for the traditional rotor initial position identification method,speed and reliability is hard to guarantee under zero and low-speed conditions,a reliable rotor initial position identification method based on composite signal injection is proposed in this paper.The HF square-wave voltage injection method cannot distinguish the polarity of permanent magnet(N pole/S pole),since two convergence points appear in the initial position identification.The polarity identification is achieved by comparing the magnitude of HF response current in the d-axis after injecting low-frequency sinusoidal current into the estimated d-axis on the basis of interrupting the square-wave voltage injection.The experimental results show that the proposed identification method can distinguish between the HF response currents in the d-axis effectively for both N and S poles of the permanent magnet to ensure the reliable identification of polarity.Meanwhile,the proposed method can estimate the initial rotor position rapidly at zero and low speed.Secondly,in order to address the problem that the system delay affects the rotor position detection accuracy and system dynamic performance when the PMSM operates at low switching frequency under low-speed condition.A gradient descent-based square wave voltage injection delay compensation method is proposed in this paper.The current loop stability is analyzed using the D-partition method,and the relationship between the system delay,the stability region of the current loop controller and the PMSM system stability is revealed.The gradient descent delay compensation method is investigated to compensate the lagged phase of the position demodulation signal and the estimated position in the square wave voltage injection method.The experimental results show that the proposed method has a wider compensation range than the traditional voltage delay compensation method,which can compensate the sampling delay,PWM update and calculation processing delay,etc,leads to the improvement in the position estimation accuracy,and enhances the steady-state and dynamic performance in the PMSM sensorless control with load condition.Again,considering theeffect of the inverter nonlinearity and magnetic field spatial harmonics,the 6k error pulsations appears in model-based method rotor position estimation,and digital delay will cause da bias errors,an improved phase-locked loop(IPLL)-based a sliding-mode observer(SMO)rotor position estimation method is proposed in this paper.The conventional SMO-PLL cannot eliminate the position error pulsation and bias error.The proposed method optimizes the conventional PLL in the perspective of input and feedback,and uses Levenberg-Marquardt(LM)to filter the specified harmonic components in the position error after pre-compensating the lag angle in the back electromotive force(backEMF)by linear prediction.Experimental results reveal that the proposed method can effectively suppress the 6th harmonic pulsation and bias error in the estimated rotor position,and improve both speed transient and load transient performance of the PMSM.The proposed method has good generality and can be easily extended to other rotor position detection methods.Finally,the composite algorithm is hard to switch smoothly in the transition zone of PMSM full speed range under load disturbance,a speed-weighted fusion method considering disturbance compensation is proposed in this paper.A feedback gain-based speed weighted fusion structure is designed to fuse the rotor position and speed estimated by two algorithms,the HF signal injection method and the sliding mode observer.A disturbance compensation method based on the q-axis feedback current is discussed to improve the stiffness of the PMSM control system,since the estimation accuracy of the fusion structure is affected by the load variation.Experimental results show that the proposed fusion method can achieve smooth switching of the composite algorithm in the range of 5%-10%rated speed,and the PMSM system behaves good anti-disturbance performance under sudden load changes.