Complex-Coefficient Adaptive Extended State Observer for Position Estimation of PMSMs With Enhanced Robustness to Disturbances

The conventional linear extended state observer (LESO) for estimating the back electromotive force (BEMF) of permanent magnet synchronous motors (PMSMs) exhibits a second-order low-pass characteristic, which leads to steady-state position estimation errors. Besides, due to the disturbances of dc errors and high-order harmonics, the conventional LESO cannot achieve high robustness to the disturbances without the help of additional digital filters, which also lead to harmonic position estimation errors. To reduce the position estimation errors, a complex-coefficient adaptive extended state observer (CCF-AESO) is proposed. In the proposed observer, a second-order adaptive complex vector filter (ACVF) with enhanced robustness to the dc errors and high-order harmonics is predesigned, and the gains of the proposed observer are configured based on the predesigned transfer function to achieve the same frequency characteristics as the ACVF. Compared with the conventional LESOs, the proposed CCF-AESO can eliminate the steady-state position estimation error and has smaller harmonic position estimation error. Comparative experiments conducted on a three-phase PMSM platform verify the proposed observer can achieve excellent position estimation even when the speed decreases to 10% of the rated speed.