A SOH Estimation Method for Lithium-Ion Batteries Based on Transfer Learning

As electric vehiclesz continue to evolve rapidly, Lithium-ion batteries become a crucial and irreplaceable power source in this industry. Accurately estimating the State of Health (SOH) is vital for the safety of it. A novel method is proposed in this paper, which combines Bi-GRU neural networks and transfer learning(TL). Key features such as Differential thermic voltage (DTV) and Internal Capacities (IC) are extracted and fused to enhance SOH estimation. In addition, transfer learning techniques are applied for the battery data from different types across different datasets, enabling the model effectively using for the new prediction tasks and enhance its transferability. The evaluation of the proposed method is carried out using public datasets, with an enhancement of accuracy by 45.7% and performance of improvement by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{7 3. 0 \%}$</tex> of different battery types. The results confirm that this method surpasses traditional estimation techniques in both accuracy and robustness, effectively enhancing the predictive capability for SOH of Lithium-ion batteries.