A multi-stage approach for energy management in microgrids

Abstract

Renewable energy sources have gained popularity over the years due to their numerous benefits, including a positive environmental and economic impact. However, stakeholders involved in electric microgrids face complex decisions when managing energy generation and consumption, particularly given the uncertainty of weather conditions. This study aims to optimize decision-making processes in energy management systems for microgrids connected to a main grid while accounting for stochasticity and dynamism in energy generation and consumption by microgrid agents. The study incorporates stochastic energy prices for purchases and sales to the main grid to model real-world uncertainties accurately. This approach enables the representation of the unpredictable nature of energy markets, which can significantly affect the performance of microgrids. The study employs the Stochastic Dual Dynamic Programming (SDDP) algorithm on a multi-stage adaptation of the microgrid model to evaluate its effectiveness compared to a deterministic equivalent model. The SDDP reduces by 11.25%, 9.45%, and 4.93% the total cost when using an instance with 32, 100, and 150 prosumers, respectively, over a 24 hours planning horizon. Overall, the results of this study have significant implications for energy management systems in microgrids, particularly in ensuring optimal performance and reducing costs. The findings can also help to take policy decisions related to renewable energy and microgrid development, particularly considering growing concerns around climate change and the need to transition towards a more sustainable energy system.