| Permanent magnet synchronous motor(PMSM) has been widely used in industry,transportation, aeronautic and astronautic fields for its high efficiency, high power density and excellent dynamic performance. More and more attention has been paid on the performance improvement and sensorless control of PMSM drives in recent years. The offline parameter identification of PMSM is essential for the controller and the observer design as they are very sensitive to parameter variations. Due to the negative effects of inverter nonlinearities, the induced current contains obvious harmonic components which cause identification error when adopting high frequency(HF) injection method for identifying d-q axis inductances. Furthermore,the existence of inverter nonlinearity can also result in pulsation of estimated rotor position in sensorless control PMSM drive based on HF square wave injection approach, which leads to inaccuracy in the coordinate transformation and result in performance degradation of the drive system. In order to improve the performance of the PMSM control system, this thesis focuses on the analysis and compensation methods of inverter nonlinearities considering HF signal injection.Firstly, this thesis investigates the inverter nonlinearity negative effects considering harmonics in the induced HF current and analyzes the hybrid strategy of combining direct current(DC) offset and HF voltage injection. By analyzing the mathematical model of the PMSM considering the influence of inverter nonlinearities, a standstill offline parameter identification method based on HF sinusoidal voltage injection is proposed. In addition, the amplitude of HF signal is calculated by discrete Fourier transform(DFT), so the estimation values of dq-axis inductances can be got by analyzing the relationship between the voltage and current amplitudes.Then, the second and third harmonics of induced current caused by the inverter nonlinearities under different injection conditions is researched and the negative effects on inductance identification are also investigated. In order to reduce the parameter estimation error, an adaptive compensation method is proposed after the self-learning process of the inverter nonlinearities, which can attenuate the negative effects of the inverter nonlinearities.On the base of senseless control of PMSM by using HF square wave injection,the envelope extraction method of induced HF current which contain rotor position information is studied. In addition, the pulsation error on rotor position estimation caused by the inverter nonlinearities will be investigated. The different error distributions caused by fundamental and HF components is analyzed. Afterdescribing the zero-current clamping effects, the negative influence of the inverter nonlinearities will be restrained by improving the injection method from HF signal amplitude regulation and signal extraction strategy.Based on the theoretical analysis, the simulation with Matlab/Simulink was given to verify the feasibility of the proposed offline parameter identification and sensorless control strategies. Finally, the effectiveness of the proposed strategies is verified with the experimental results at a 2.2k W PMSM sensorless drive. |
PDF Full Text Request