FPGA-Implementation Friendly Long-Horizon Finite Control Set Model Predictive Control for High-Power Electronic Systems

In this paper we present a branch-and-bound algorithm that facilitates the real-time implementation of long-horizon finite control set model predictive control (FCS-MPC) for high-power electronic systems. To achieve this, the optimization process is modified such that the more likely candidates (i.e., the switching sequences with less switching transitions) are explored first. By doing so, computational demanding operations can be efficiently pipelined in a field-programmable gate array (FPGA), rendering the proposed method computationally feasible. The effectiveness of the introduced method is tested both in an FPGA-in-the-loop simulation and with a real-world setup based on a scaled-down low-voltage variable speed drive system consisting of a three-level neutral point clamped inverter and an induction machine. As shown, a real-time implementation with horizon length of twelve is achieved without sacrificing optimality.