An enhanced deep learning framework for state of health and remaining useful life prediction of lithium-ion battery based on discharge fragments

Accurate estimation of state of health (SOH) and remaining useful life (RUL) is crucial for enhancing the reliability and safety of battery systems. The aging mechanism of lithium-ion batteries is complex and exhibits substantial nonlinearity, posing significant challenges for achieving precise estimation. This study proposes a capable deep neural network framework integrates deep convolutional neural network, long short-term memory network, and multi-head attention mechanism. Accurate prediction can be achieved by only utilizing random voltage-current data segments from discharge process with the proposed framework. The experimental results demonstrate the effectiveness of the proposed method, the root mean square error for SOH and RUL estimation are 0.1751 % and 0.90 in the NASA dataset and 0.1584 % and 13.78 in the experimental dataset, respectively. The proposed method can accomplish accurate prediction without complex feature extraction, significantly enhancing its practical applicability.