Real-time Implementation of Model Predictive Space Vector Modulation for a Three-Level Neutral-Point-Clamped Inverter

Abstract

The goal of this paper is to propose a resource-efficient, real-time implementation architecture for a new inverter modulation scheme called model predictive space vector modulation (MPSVM), which embeds the solution of an optimization problem in the modulation algorithm. MPSVM comes with a high computational burden. Firstly, this paper introduces computational methods to decrease the computational Complexity of MPSVM and solve it more efficiently. Second, The real-time implementation is done on an FPGA (field-programmable gate array), to achieve computational speeds In just a couple of microseconds. Furthermore, FPGA-in-the-loop (FIL) simulations are performed to validate the FPGA design. Finally, this paper shows that the architecture proposed is resource-efficient and can achieve computational speeds at the microsecond level, and possible extensions for future work are provided.

Date of Award	21 Jun 2023
Original language	English
Supervisor	George Papafotiou (Supervisor 1) & Victor T.T. Lam (Supervisor 2)