A Novel Comparative Analysis of Deep Learning for Uncertainty Management in Networked Microgrids

Increased carbon emissions and energy demand have risen due to population growth and higher consumption. Networked microgirds (NMGs) integrated with renewable energy resources (RERs) offers a sustainable solution, but the intermittent nature of solar and wind power challenges grid stability and reliability. Accurate energy forecasting is essential for efficient energy management and optimal resource utilization. Deep learning (DL) and machine learning (ML) have emerged as effective tools for forecasting renewable energy generation and demand in NMGs. However, a fair, quantitative comparison of DL-based forecasting approaches is lacking in the existing literature. This article proposes a novel approach for comparing DL-based forecasting approaches for solar, wind, and load prediction based on performance score. Results show that forecasting accuracy depends on overall performance score of forecasting approach. Finally, we highlight key challenges and future opportunities, focusing on better data management, improved prediction accuracy, and scalable AI solutions for uncertainty forecasting in networked microgrids. This study serves as a valuable resource for advancing DL-based forecasting in sustainable energy management in NMGs.