Hybrid Active Flux Observer to Suppress Position Estimation Error for Sensorless IPMSM Drives
Qilian Lin, Ling Liu, Deliang Liang
Abstract
Harmonics and drifts are found in the position estimation errors of high-frequency signal injection (HFI) methods. To suppress the position estimation error, a hybrid active flux observer with disturbance rejection for both HFI method and model-based method is proposed. It fuses the position information from traditional HFI method and model-based method as the input of a disturbance observer, the estimated disturbance output can compensate measurement error for the two methods. Then a reduced-order natural speed observer is connected in a cascaded form to reject motor mechanical disturbance by estimating the load torque. Besides, a stator resistance identification algorithm is proposed, which is proved stable near zero frequency under persistent excitation condition. A tuning guideline for the whole algorithm is elaborated based on a rigorous stability analysis. Effective experiments and simulations verify the feasibility of the proposed scheme on a 0.75-kW interior permanent magnet synchronous motor drive platform.