Model Predictive Speed Control With QGI-ESO for Suppressing Speed Periodic Disturbances of SPMSM

The model-free predictive control based on the extended state observer (ESO) has limitations in suppressing unmodeled periodic disturbances. Therefore, this article proposes a model predictive speed control (MPSC) method that embeds a quasi-generalized integrator in an extended state observer (QGI-ESO). First, a speed ultralocal model is constructed to analyze the control structure of lumped disturbances, such as periodic disturbances arising from current measurement errors and inverter dead time. Second, according to the harmonic characteristics, a QGI-ESO for suppressing the 1st and 2nd harmonics is designed in the speed loop, and a QGI-ESO for suppressing the 6th and 12th harmonics is designed in the current loop. Subsequently, the observer’s stability is proved, and its frequency characteristics are analyzed. Meanwhile, the sensitivity functions for high-frequency disturbances and speed control are derived to confirm the observer’s suppression of the target disturbances. Finally, this control strategy is applied to the surface-mounted permanent magnet synchronous motor (SPMSM) experimental platform, based on the DSP F28379D, to verify its advantage in reducing speed ripple.