A Flexible State-of-Health Prediction Scheme for Lithium-Ion Battery Packs With Long Short-Term Memory Network and Transfer Learning
Xing Shu,Jiangwei Shen,Guang Li,Yuanjian Zhang,Zheng Chen,Yonggang Liu
DOI: https://doi.org/10.1109/tte.2021.3074638
IF: 6.519
2021-12-01
IEEE Transactions on Transportation Electrification
Abstract:The application of machine learning-based state-of-health (SOH) prediction is hindered by the large demand for training data. To conquer this defect, a flexible and easily transferred SOH prediction scheme for lithium-ion battery packs is developed. First, the charging duration for a predefined voltage range is hired as the health feature to quantify capacity degradation. Then, the long short-term memory (LSTM) network and transfer learning (TL) with fine-tuning strategy are incorporated to constitute the cell mean model (CMM) for SOH prediction with partial training data. Next, to evaluate the SOH inconsistencies among cells, the LSTM model is employed as the cell difference model (CDM), and the minimum estimation value of CDM is identified to determine pack SOH. The experimental results reveal that, even when the first 360 cycle data, occupying only 40% in the whole 904 cycle data, are chosen and constituted to the data set for model training, the obtained estimation algorithm can still predict SOH precisely with the error of less than 3%, thus remarkably reducing the training data amount and mitigating the computation burden during model training. In addition, the preferable validation results on different types of lithium-ion batteries further manifest the extendibility of the proposed strategy.
engineering, electrical & electronic,transportation science & technology