Research On Position Sensorless Control In Medium-high Speed And Starting Strategy Of Permanent Magnet Synchronous Machine

Permanent magnet synchronous machines(PMSM)position sensorless control technology uses position detection algorithms to replace mechanical encoders,which can effectively reduce system costs,reduce equipment size,and improve reliability.It has remarkable application value and research significance.At present,the position sensorless speed control system still has technical problems such as slow dynamic performance of position estimation and poor starting process stability.This dissertation focuses on the position sensorless control technology in the middle-high speed range and the starting strategy of the motor in different initial states,aiming to improve the working efficiency and starting stability of the speed control system.The flux observation scheme needs to adopt the filter and phase-locked loop structure to realize position and velocity information tracking.Aiming at the position observation deviation introduced by the filter and the phase-locked loop,this dissertation proposes a position compensation strategy based on the linear extended state observer(LESO).With LESO constructed in the synchronous rotating coordinate system with a deviation of orientation angle,real-time observation and feedforward compensation of position deviation are carried out.Furthermore,in view of the static deviation of LESO's observation of slope disturbance,PI feedback is introduced into the observer structure to improve the observation accuracy during motor acceleration and deceleration.The proposed compensation strategy improves the position observation accuracy,improves the dynamic response performance of the position observer,and further improves the working efficiency of the drive system.In order to effectively utilize the salient pole effect of the motor and improve the working efficiency of the motor during steady-state operation,this dissertation focuses on the control method of the maximum torque per ampere(MTPA).Based on the analysis of the traditional virtual sine signal injection scheme,it is pointed out that the filters used in series in the signal processing process causes the dynamic tracking performance of the system to decrease,and a MTPA control strategy based on composite virtual signal injection is further proposed.By selecting multiple injection signals with appropriate phases,the filters in the process of obtaining effective information is eliminated,thereby improving the dynamic tracking performance of the system.When the motor starts from the zero/low-speed state,the flux observation scheme based on the voltage model is difficult to obtain the accurate rotor position when the motor is running at low speed,so it is necessary to combine the low-speed control scheme based on the salient pole effect.Aiming at the problem of easy oscillation during switching due to the structural differences between the two schemes,this dissertation proposes a switching strategy based on the virtual q-axis inductance.First,the internal connection between the position observation value obtained by the flux linkage observation method and the motor parameters in the flux linkage calculation model is analyzed,and then the virtual q-axis inductance is constructed to make the two schemes obtain the same position observation value.According to the relationship between the virtual q-axis inductance value and the position deviation of the two solutions,a control strategy for online adjustment of the virtual q-axis inductance value is proposed.The proposed strategy realizes the fast and smooth switching of the two schemes.When the motor is in a high-speed rotation state before starting,the detection of the initial position of the rotor is easy to produce an inrush current that affects the smooth start of the motor.To solve this problem,this dissertation proposes a zero-voltage vector scheme based on double closed-loop control of the short-circuit current amplitude.On the one hand,a closed-loop control system on the amplitude of the short-circuit current is constructed so that the zero-voltage vector action time can be automatically adjusted when the motor is at different initial speeds.On the other hand,analyze the working conditions of the zerovoltage vector method,and design a closed-loop control strategy for the set value of the short-circuit current amplitude on this basis.Finally,the phase-locked loop algorithm is used to obtain the initial position and speed information of the rotor from the short-circuit current,which improves the observation accuracy compared with the traditional scheme.This dissertation focuses on the position sensorless control technology in the highspeed range of the permanent magnet synchronous motor and the starting strategy under different initial states.The actual test platform experiments verify the feasibility and effectiveness of the proposed control schemes.