Inertia Identification in Power Systems with Integrated Renewable Energy Using WRARMAX-LSTM

To effectively track the inertia of a power grid, two main objectives are prioritized: reducing computational time and improving identification accuracy. This study employs the WRARMAX model (Weighted Recursive Least Squares solving Autoregressive Moving Average model with exogenous input) and leverages multithreaded parallelism to significantly decrease identification time. Additionally, we propose using a Long Short-Term Memory (LSTM) network to process the residuals from the WRARMAX model. This hybrid approach aims to enhance the accuracy and stability of inertia identification, addressing the dynamic challenges posed by the increasing integration of renewable energy. By combining these methodologies, we present a robust and efficient framework for real-time inertia tracking in modern power systems.