Gerenciamento Inteligente de microrredes: impacto da modelagem matemática na eficiência computacional de algoritmos de otimização

Abstract

A microgrid can be defined as a localized group of loads and energy sources, typically relying on renewable energy as its primary generation source, along with an energy storage system. Microgrids can operate independently from the traditional power grid due to factors such as the need to supply electricity to remote areas, enhancing system resilience, and reducing dependency on centralized infrastructures. Microgrid management involves determining the optimal energy dispatch for each microgrid element, ensuring demand fulfillment while minimizing operational costs and pollutant emissions. This management process is often carried out using intelligent management systems that employ artificial intelligence-based algorithms to optimize energy resource allocation. A widely used methodology in this context is Model Predictive Control (MPC), which integrates forecasting and optimization algorithms for energy management. One of the key aspects of optimization algorithms lies in their mathematical formulation. An inadequate formulation can lead to higher computational costs and inefficient solutions. This thesis investigates and proposes improvements in the mathematical formulation of optimization algorithms for intelligent management of isolated microgrids, focusing on the relationship between operational costs, pollutant emissions, and Operation and Maintenance (O&M) costs of non-dispatchable generators. Mono-objective and multi-objective approaches were analyzed, comparing existing methodologies with innovative proposals aimed at reducing computational complexity and improving the quality of obtained solutions. The tests conducted under different scenarios and configurations—including cases with a single pollutant-based generator and multiple sources—demonstrated that the improved formulation is more efficient and provides a more robust balance between efficiency and solution diversity, contributing to the economic viability and environmental sustainability of isolated microgrid management.