Energy Management for an AC Island Microgrid Using Dynamic Programming

This paper presents an optimal Energy Management Strategy for distributed generations in AC island microgrid. The considered microgrid consists of PV generator, wind turbine, gas turbine, battery, supercapacitor, load, electric vehicle and hydrogen energy storage system consisting of fuel cell, electrolyzer and hydrogen tank. A centralized control has been used in this work to optimize the energy management while ensuring a power balance between production and consumption as well as optimal operating conditions. Two predictive energy control strategies, i.e. requiring predictive data, based on dynamic programming and rule-based strategy have been proposed for this topology. The main goal is to minimize the operating cost of the system and to maintain the balance between generation and consumption as well as optimal operating conditions. Indeed, some input data are difficult to predict with a good reliability. To cope with these prediction errors, so-called "reactive" algorithms are developed, capable of modifying the predictive strategy according to the prediction errors and aiming at predetermining the production profile of the generators, in order to achieve a comprehensive optimization of an objective function for the power system and subsequently adjust the operating points during the day. Simulations are performed to test the performance of the methods proposed.