PMSM Speed Ripple Suppression Due to Current Measurement Error Using Quasi-Fractional Resonant-Normalized Extended State Observer

Abstract

Current measurement error causes periodic speed ripple in permanent-magnet synchronous motor control systems. The typical normalized extended state observer (NESO) can estimate and compensate for the load disturbance but is not sufficiently capable to deal with the periodic disturbances. Thus, a speed controller based on the NESO frame, combining a PD controller and a quasi-fractional-resonant (QFR) controller, is proposed to address this speed ripple issue due to the current measurement error including offset error and scaling error. Then, it is verified that the proposed controller satisfies the separation principle and QFR controller does not affect the open-loop characteristics of the system through mathematical derivation in the frequency domain. Besides, the proposed controller in outer speed loop and a PI controller in inner current loop are designed with the proposed analytical parameters’ tuning methods based on frequency-domain analysis. Compared with the existing adaptive proportional-integral–resonant (PIR) controller, the proposed controller not only suppresses the speed ripple to a lower level but also achieves better speed tracking and load disturbance rejection performances. The experimental result comparison confirms the effectiveness of the proposed controller and design scheme.