Research On Adaptive Compensation Methods Of Rotor Position Estimation Error For PMSM Drives

Interior permanent magnet synchronous (IPMSM) motor has been widely usedin many drive fields for its high efficiency, high power density, strong robustness,excellent speed control performance and the easiness of flux-weakening control.However, the installation of the mechanical position sensor increases the cost anddecreases the reliability of the IPMSM drive system. Consequently, various positionsensorless control methods have been getting increasing attention. Adopting theelectromotive force (EMF) based methods, the factors that parameter variation, fluxspatial harmonics and inverter nonlinearities will result in substantial harmonicerror existing in position estimation. However, in the vector controlled IPMSMdrive system, the harmonics existing in position estimation will lea d to inaccuracyin the coordinate transformation, and cause the torque ripple, resulting in theperformance degradation of the drives. This thesis focuses on the adaptivecompensation methods of IPMSM rotor position estimation error to improve theperformance of the senseless IPMSM drive system.The IPMSM rotor position observation method based on extended EMF modelis established. Based on the mathematic model of IPMSM and the principle ofvector control, the extended EMF observer model has been built. Combining theorthogonal software phase lock loop (PLL), the position observer has been designed.The relationship among the phase lag, harmonic suppression of the extended EMFobserver and operating frequency is established. In addition, the mechanism ofbeing caused rotor position estimation error by the inverter nonlinearity, magneticsaturation and spatial harmonic is analyzed.A dead-time compensation strategy based on adaptive trapezoidal error voltageis studied. Since the inverter nonlinearity will lead to the substantial harmoniccomponent existing in the phase current and estimated EMF. And then make theestimated position contain harmonic error, which influences the performance of thesensorless control system. Based on the analysis of distorted voltage caused byinverter nonlinearity, the adaptive regulating mechanism of trapezoidal errorcompensation voltage has been studied. According to the operating condition, thetrapezoidal angle of the compensation voltage can be changed adaptively, achievingthe real-time compensation for the inverter nonlinearity.To further improve the performance of sensorless IPMSM drives, an adaptiverotor position harmonic error suppression method based on second-ordergeneralized integrator (SOGI) adaptive filter is studied. A harmonic decouplingnetwork consisting of multiple SOGIs is adopted to achieve the multiple selective EMF harmonic elimination (MSEHE). The application of the frequency-locked loopensures the SOGI resonance frequency adaptive. The research method can eliminatethe low-order harmonic component effectively, suppressing the position estimationharmonic error.Based on the theoretical analysis, simulation with Matlab/Simulink was givento verify the feasibility of the proposed control strategies. The effectiveness of theproposed strategies is verified with the experimental results at a2.2kW IPMSMsensorless drive.