AN MPC-BASED ENERGY MANAGEMENT SYSTEM EMBBEDING NEURAL NETWORKS FOR FORECASTING RENEWABLE ENERGY GENERATION AND DEMAND: A STUDY ON GREEN HYDROGEN PRODUCTION

This study proposes a Model Predictive Control (MPC) framework for microgrid management, using Artificial Neural Networks (ANNs) to forecast renewable energy generation and load demand. These forecasts are integrated into a two-layer MPC strategy based on quadratic programming to optimize real-time control and address renewable source variability and demand uncertainty. A case study in Guanambi, Bahia (Brazil) demonstrates the method's effectiveness in improving energy efficiency and enabling green hydrogen production through electrolysis. The results highlight the potential of the MPC-ANN framework to drive sustainable energy transitions and industrial decarbonization in distributed energy systems.