Experimental Analysis of K-Best Sphere Decoding Algorithm for LPH-FCS-MPC

Abstract

Finite control set model predictive control (FCS-MPC) with long prediction horizon (LPH-FCS-MPC) is becoming a great option for the control of power conversion systems. In contrast to short prediction horizon FCS-MPC, extending the prediction horizon length renders important steady-state improvements as the harmonic distortion and/or the switching frequency can be considerably reduced, for a more efficient operation of the power converter. The main disadvantage is that practical application of these strategies is still limited due to computational costs. In this work, a recently proposed hardware K-best sphere decoding algorithm (K-best SDA) is further analyzed and verified through experiments in an uninterruptible power supply (UPS) system. The results indicate the efficiency of the algorithm, enabling longer prediction horizons and better control performance in comparison with the conventional SDA.