Supervised Imitation Learning of Finite-Set Model Predictive Control Systems for Power Electronics

In the past years, finite-set model predictive control (FS-MPC) has received a lot of attention in power electronics field. Due to very simple inclusion objectives and straightforward design, it been adopted different converter topologies. However, computational burden often imposes limitations implementation if multistep predictions are deployed or/and multilevel converters with many possible switching states used. To remove these limitations, we propose imitate controller. It is important highlight that imitator not intended improve dynamic or steady-state performance original FS-MPC algorithm. contrast, its key role keep approximately same while at time reducing burden. Our proposed an artificial neural network trained offline using data labeled by Since correlated complexity algorithm emulates, much more complex controllers made without prior limitations. The method validated experimentally on stand-alone configuration results confirm good match between controller performance. Simulation models both provided supplementary files for three prediction horizons.