Sequential Direct Model Predictive Control for Gird-Tied Three-Level NPC Power Converters

A sequential direct model predictive control (SDMPC) for the grid-tied three-level neutral-point clamped (3L-NPC) converters is proposed in this work. For the underlying 3L-NPC grid-tied power converter, control objectives are active/reactive current tracking and neutral-point voltage balance. The proposed solution utilized two cascaded cost functions, dealing with the above control objectives in a sequential manner without using weighting factors, such that the tedious parameter tuning process is eliminated. It is different from the classical direct model predictive control (CDMPC), which combines both control objectives into a single cost function with a tunable penalty coefficient, i.e., the weighting factor. A thorough comparison investigation between CDMPC and the proposed method was carried out. The simulation results showed that both methods achieve comparably good performance. While SDMPC considerably reduces the tuning efforts, and simplifies the engineering design process.