Robust Predictive Control of Three-Level NPC Back-to-Back Power Converter PMSG Wind Turbine Systems With Revised Predictions

A direct-drive permanent magnet synchronous generator (PMSG) with a
three-level
neutral-point-clamped back-to-back power converter is an attractive configuration for
high-power
wind energy conversion systems. For such a topology, finite-control-set model-predictive control (FCS-MPC) has emerged as a promising alternative. However, due to its fully model-based concept, variation of system parameters (in particular, the stator and grid filter inductance and rotor permanent-flux linkage) will (seriously) affect the system control performances when using the classical FCS-MPC. In this work, a
robust
FCS-MPC method with
revised predictions
is proposed and validated for such a system. With the proposed solution,
not only
the system robustness against parameter variations is improved,
but also
the control variable ripples are evidently reduced. The proposed method has been implemented with a fully field-programmable-gate-array-based real-time hardware. Its performance improvements in comparison with the conventional solutions are validated with
experimental
data.