Digital Twin-Based Model of Battery Energy Storage Systems for SOC Evaluation

• Published in: 2023 3rd Power System and Green Energy Conference (PSGEC) pp. 1075 - 1079
• Main Authors: Sun, Yushu, Pu, Xiaowei, Xiao, Hao, Han, Daoxin, Li, Yixuan, Guo, Zhiming, Zhao, Jian
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2023
• Subjects: Batteries; battery energy storage systems; Bi-LSTM; digital twin; Digital twins; Lithium batteries; neural network fusion method; Power system reliability; Prediction methods; Recurrent neural networks; RNN; SOC; Uncertainty
• DOI: 10.1109/PSGEC58411.2023.10255915

The battery energy storage system is a complex and non-linear multi-parameter system, where uncertainties of key parameters and variations in individual batteries seriously affect the reliability, safety and efficiency of the system. To address this issue, a digital twin-based SOC evaluation method for battery energy storage systems is proposed in this paper. This method enables accurate state estimation of the SOC, mitigates safety hazards and prevents unforeseen system failures. By utilizing digital twin technology, Recurrent Neural Network (RNN) and Bi-Long Short-Term Memory (Bi-LSTM) neural network are integrated to develop multiple neural network fusion methods, and the split-band prediction is further used to improve SOC prediction accuracy. The proposed method effectively overcomes the limitations of single prediction method for different research subjects, which can provide decision support for accurate SOC assessment and ensure optimal performance and safe operation of the battery energy storage system.