Modified Model Predictive Control of Back-to-Back T-type NPC Converter Interfacing Wind Turbine-Driven PMSG and Electric Grid

Abstract

In this paper, a modified model predictive control (MMPC) scheme is proposed for controlling a back-to-back (BTB) T-type neutral-point-clamped converter (TNPCC). The BTB TNPCC ties a permanent magnet synchronous generator (PMSG) to an electric grid. The PMSG is driven by a variable speed wind turbine. The aim of the proposed method is to mitigate the CMV not only on wind generator-side and but also on grid-side of BTB TNPCC and to reduce the computation effort. Classical MPC (CMPC) algorithms CMPC-I/CMPC-II and MMPC algorithms MMPC-I/MMPC-II are modeled, implemented and compared for two cases: zero CMV on GS/WGS of BTB TNPCC. The simulation results confirm that the MMPC scheme shows similar or slightly improved performance than that of CMPC scheme while accomplishing the following control goals: maximum power point tracking, reactive power control, DC-link voltage control, DC-link capacitors voltage balancing, and CMV mitigation. Moreover, the MMPC strategy shows additional advantages: It reduces a) the computational effort b) the peak CMV on the GS and keeps zero CMV on the WGS, or vice versa. Sensitivity tests are performed to analyze the performance of proposed algorithms in a complex scenario, where both GS and WGS inductance parameters are underestimated/overestimated concurrently. The MMPC strategy has been validated experimentally on an FPGA controller.