LSTM-Based Digital Twin Modeling for LVDC Microgrid Operation with Pulsed Load
Conference
Abdelrahman, MS, Esoofally, M, Hussein, H et al. (2024). LSTM-Based Digital Twin Modeling for LVDC Microgrid Operation with Pulsed Load
. 10.1109/IAS55788.2024.11023698
Abdelrahman, MS, Esoofally, M, Hussein, H et al. (2024). LSTM-Based Digital Twin Modeling for LVDC Microgrid Operation with Pulsed Load
. 10.1109/IAS55788.2024.11023698
Proper control and energy management is crucial for the effective, practical, and reliable operation of DC microgrids. In DC microgrids, the high certainties of renewables and the dynamic behavior of particular loads propose different modeling, operation, and control challenges. With recent advancements in information and communication technologies offering energy solution services relying on data analytics, machine learning (ML), and IoT technologies, the concept of digital twins (DT) has emerged as a feasible and active model to enhance the operation and resilience of microgrids. This paper proposes a long short-term memory (LSTM) neural networkbased strategy to construct a data-driven digital twin (DT) model of DC microgrids. The model is tested and validated with a standalone DC MG containing Renewable Energy Sources (RES), Energy Storage Sources (ESSs), and different load types. The accuracy of the created twin model is tested and evaluated under different scenarios of the microgrid operation, such as a sudden decrease in renewable generation and energizing pulsed load to the MG.
