Analysis and prediction of battery aging modes based on transfer learning

Aging modes analysis of lithium-ion batteries plays a crucial role in battery health management. The present studies for battery aging modes analysis are mainly based on mechanistic models or electrochemical models. However, most of the parameters of these models need to be measured offline, which adds difficulties to actual vehicle applications. Therefore, this paper proposes a method for analyzing and predicting battery aging modes based on a transfer learning method. Aging modes data of experimental batteries and electric vehicle batteries (EVBs) are obtained by an enhanced dual-tank model. Then, a transfer learning model based on long and short-term memory neural network is trained to achieve the analysis of EVBs aging modes by using the experimental battery aging data as the source data and the aging data of EVBs as the target data. The results show that the estimation error for the EVBs aging parameters is less than 6% and the predicted error for the next 5 months is less than 4%. The method offers a new idea for EVBs online aging modes analysis and prediction.