An improved model predictive control of back-to-back three-level NPC converters with virtual space vectors for high power PMSG-based wind energy conversion systems.

In this study, an improved virtual space vector (VSV)-based two-stage model predictive control (MPC) scheme is presented for neutral point clamped (NPC) converters in high power-rated permanent magnet synchronous generator (PMSG)-based wind energy conversion system applications. The presented MPC scheme utilizes the VSVs and involves two-stage prediction for effectively reducing computation complexity. In stage-I prediction, a virtual space vector is identified based on the optimum cost function. Then, a set of sub-sector vectors are placed in stage II prediction based on the capacitor voltage levels. From this, the cost function predicts the optimum switching state. At the same time, the conditional selection of small voltage vectors in stage II prediction eliminates the neutral point voltage balancing constraint. Finally, the presented method is verified by the simulation of the back-to-back NPC converter of a high-power-rated PMSG-based wind energy conversion system. Further, the adaptability of the presented scheme for maximum power point tracking, reactive power control, and dc-link voltage control is also verified by simulation.