Abstract：Model predictive control（MPC）is an effective control strategy for permanent magnet synchronous

generators（PMSG）due to its fast dynamic response and multi-objective optimization capabilities. However，MPC

relies on accurate system models and sensor measurements. In practical conditions，parameter mismatch caused by

PMSG parameter variations and sensor measurement noise can deteriorate the control performance of MPC. Robust

predictive control based on extended state observer （ESO） can effectively deal with parameter mismatch.

However，a single-gain ESO is difficult to balance parameter mismatch and measurement noise disturbance.

Therefore，a robust predictive control method based on hybrid cascade parallel ESO（CPESO）was proposed，which used multiple sub-ESOs in series and parallel to weight system disturbances and observed values for noise

suppression. This method can effectively balance parameter mismatch and measurement noise suppression. Finally，

under conditions with parameter mismatch and measurement noise，experiments were conducted on a three-level

PMSG test bench to verify the effectiveness of the proposed method.

      Key words：permanent magnet synchronous generator（PMSG）；model predictive control（MPC）；noise

suppression；disturbance rejection；robust control