Integrated Digital Twin Design and Prediction-Aided Control of PMSG-Powered Sodium-Ion Battery Systems

This paper presents an enhanced Digital Twin (DT) framework that integrates a Permanent Magnet Synchronous Generator (PMSG) with a Sodium-Ion Battery (SIB) storage system. Unlike conventional DT-based Battery Management Systems (BMS), the proposed architecture incorporates a hybrid State-of-Charge (SoC) and State-of-Health (SoH) estimation scheme, capacity-fade-driven SoH modeling, and Long Short-Term Memory (LSTM) based performance prediction tailored for SIB chemistry. A novel coupling of LSTM forecasting with Model Predictive Control (MPC) is introduced to optimize the charge-discharge scheduling under fluctuating wind energy inputs. Simulation results demonstrate improved SoC estimation accuracy ($>96 \%$), reduced LSTM forecasting error, superior MPC performance compared to a baseline PI controller, and an estimated battery lifetime enhancement of approximately 14 %. The results validate the feasibility of real-time DT updating and prediction-assisted energy control for next-generation renewable energy systems.