Inertia Forecasting for Resilient High-IBR Power Systems using Deep Learning Techniques

The increasing integration of renewable energy sources (RES), particularly inverter-based resources (IBR) like wind and solar, has been reducing system inertia, posing risks to grid stability and potentially leading to cascading outages during power disturbances. This study has examined the impact of these resources on inertia trends and highlights the importance of accurate inertia forecasting for effective grid management. To improve forecasting accuracy, a bidirectional Long Short-Term Memory (Bi-LSTM) deep learning model has been proposed and was compared with a standard LSTM model using historical data from the Great Britain (GB) power system. Both models were evaluated using key validation metrics, with results indicating that while both forecast inertia effectively, the Bi-LSTM model achieves higher accuracy. This enhanced precision provides grid operators with essential insights for improved grid stability and optimized IBR integration, contributing to a more reliable and sustainable power system.