A New Variable-Frequency Optimal Direct Power Control Algorithm

This paper presents a new formulation for direct power control (DPC) with several improvements over previous DPC formulations such as an exact discrete-time expression for the predicted power variations yielded by a given inverter switching state and the inclusion of the computing delay time in the model. Based on this formulation, a variable-frequency optimal DPC algorithm is presented that selects the inverter switching state that minimizes the power error. This algorithm shows an excellent precision in estimating power variations, therefore reducing power ripple and unwanted current harmonics, while retaining the fast dynamics inherent to variable-frequency DPC. This makes this algorithm suitable for interfacing distributed generation to microgrids, where a fast and accurate power control is desirable in order to perform voltage and frequency regulation. Additionally, studies are carried out regarding the average switching frequency and the influence of the model parameters on the algorithm robustness.