Robustness enhancement of position sensorless super-high-speed PMSM for hydrogen fuel cell air compressors based on linear extended state observer

As a crucial component within hydrogen fuel cell systems, air compressors are driven by super-high-speed permanent magnet synchronous motors (SHSPMSMs), which notably emphasize on the robustness performance of position sensorless control. Conventional extended back electromotive force (EEMF)-based position sensorless SHSPMSM drives suffer deteriorated performance from parameter dependence, inherent steady-state position estimation error and scanty transient-state robustness. To address these issues, an effective linear extended state observer (LESO)-based solution is proposed in this article, improving the robustness of position sensorless SHSPMSM drives in the wide speed range. Primarily, based on the motor ultra-local model, two LESOs are utilized to estimate the internal disturbances corresponding to parameter variation terms rather than directly estimating EEMF. Desired estimation performance of EEMF can be achieved regardless of parameter mismatch. Additionally, taking the load torque as the external disturbance, another LESO is proposed to observe speed and position accurately and precisely. Theoretical analysis shows that zero steady-state position observation error exists during the acceleration and deceleration operations, and that faster dynamics can be achieved with adaptive bandwidth design of LESO. Finally, the effectiveness of the proposed method is validated on a 35 kW SHSPMSM drive system for hydrogen fuel cell air compressors with rated speed of 90 kr/min, affirming its superiority compared with conventional method featured as Luenberger observer-based EEMF estimation integrated with phase-locked loop (PLL)-based position extraction.