| Due to a series of strong points such as high efficiency,high power density,small size,and strong robustness,permanent magnet synchronous motors have been commonly used in the industrial field.However,the volume and cost of the PMSM drive system were affected by the mechanical positionsensor with high precision.Consequently,the sensorless control technology become more and more popula in recent years.In the medium and high speed field,the sliding mode observer-phase locked loop(SMO-PLL)system based on the enhanced electromotive-force(EMF)model has become one of the widely used sensorless position estimation methods due to its simple and reliable.However,the inverter nonlinearity,flux spatial harmonics and the measurement error result in the dc component and harmonics in estimated EMF,which in turn leads to pulse error in PLL when detecting the rotor position.In order to meet the needs of the PMSM sensorless drive system with high dynamic performance,this paper proposes two kinds of the rotor position estimation error suppression strategies based on the EMF model and the equivalent flux model,respectively.Firstly,the function and structure of the PMSM are introduc ed.And the vector control system of IPMSM was built by using SVPWM.The concept of EMF is derived,thereby constructing a sliding mode observer(SMO)based on the EMF.The EMF obtained by the SMO is input into the phase-locked loop(PLL),and the speed and position information are estimated.Then,the mathematical analysis of the harmonics and dc errors in EMF estimation and rotor position estimation error in PLL caused by the main interference factors such as inverter nonlinearity,flux spati al harmonics and the measurement error is given.Finally,simulation and experimental results prove the correctness of the aforementioned inferences.Secondly,taking the EMF estimation as the input signal of PLL.On the basis of SMO-PLL,the rotor position estimation error suppression strategy based the fifth-order generalized integrator(FOGI)is proposed.On the premise that the frequencylocked loop(FLL)ensures frequency tracking and locking at any time,the FOGI will extract the selected signal with no amplitude attenuation and no phase shift.So that the interference signal in estimated EMF is eliminated,thereby the pulse error in position estimation in PLL will be suppressed.The FOGI has multiple parameters,which overcomes the shortcomings of the reduced-order generalized integrator(ROGI)that the filtering accuracy and response speed cannot be improved at the same time due to the single parameter.Under the requirement of the same settling time,FOGI will attenuate more position errors,it is a reliable solution to ensure the accuracy of position estimation and rotor speed estimation under medium and high speed sensorless control systems.Thirdly,the equivalent flux estimation will be as the PLL input signal.The rotor position estimation error suppression strategy based on the third-order generalized integral flux observer(TOGIFO)is proposed based on SMO-PLL.Under the premise of frequency adaptation,the TOGIFO can be equivalent to a pure integral at the resonance frequency,so it can effecti vely get the estimated equivalent flux when inputting EMF estimation.During the integration operation,the TOGIFO completely eliminates the saturation phenomenon caused by dc input of EMF estimation,and greatly attenuates the harmonics of inputs,and obtains high-precision equivalent flux fundamental signal in real time,the rotor position estimation error and rotor speed estimation error in PLL will be suppressed.The TOGIFO is a feasible solution to enhance the precision of sensorless control and further broaden the speed application range of SMO system.Finally,on the basis of the foregoing theories,the reliability and feasibility of the two strategies suggested were proved by simulation research and hardware experiments that performed on the MATLAB/Simulink simulation software and the ds PACE experimental platform. |
PDF Full Text Request