Impedance Model Based Online Inductance Identification Method of Permanent Magnet Synchronous Motor Decoupled from Rotor Position Error
Permanent Magnet Synchronous Motor(PMSM)is widely used in many industrial fields because of its good control performance and high power density.The PMSM inductance is greatly affected by the saturation of working conditions,making online inductance identification important.Most traditional online inductance identification algorithms for PMSM are implemented based on the dq-axis model,which relies on the accurate rotor position.If a bias is caused by sensorless control or position sensor error,the accuracy of the traditional inductance identification method can not be guaranteed.This paper proposes an online identification method for PMSM inductance decoupling from the rotor position error.Firstly,the impedance characteristics of the PMSM inductance under the complex frequency domain are analyzed.The virtual-axis impedance model decoupling from the rotor position is constructed according to the inductance-changing characteristics at different rotor positions.Then,the inductance estimation strategy based on the high-frequency(HF)injection under the virtual axis is proposed.The whole HF injection is realized based on the dq-axis injection with the coordination transformation.The HF injected and responding signal is extracted by the fast Fourier transform(FFT).Finally,the dq-axis inductance identification is realized by searching the peak values of the virtual-axis inductance to avoid the influence of the rotor position error.Furthermore,the construction strategy of the virtual axis and the amplitude-frequency selection strategy of the injected signal are analyzed.A 2.2 kW interior PMSM test platform was built.The experimental results show that the proposed method can realize the inductance identification under different online operating conditions,which is immune to the rotor position error.Different working conditions are tested(from no-load to full-load at 15 Hz and 25 Hz)to verify the robustness of the method.Herein,the signal sampling and identification period by DFT is constant.Meanwhile,the inductance surfaces under different working conditions(different combinations of id and iq)are obtained.The inductance changes obviously under different working conditions.Compared with the unsaturated inductance,the variation range of Ld,q can reach up to 13.1%and 10.2%.The identification errors of Ld,q under different id,q combinations can be reduced to less than 5.1%and 3.6%,respectively,showing the accuracy of the proposed method.The conclusions can be drawn from the experimental results.(1)The proposed method based on the virtual-axis HF impedance model is independent of the rotor position,which is suitable for applications with rotor position error.(2)The accuracy of the inductance identification can be guaranteed through the construction strategy of the virtual axis and the amplitude-frequency selection strategy.
Permanent magnet synchronous motor(PMSM)online inductance identificationhigh frequency signal injectionequivalent impedance modelrotor position error decoupling
国家科技期刊平台
NSTL
万方数据
