State of health estimation for lithium-ion battery based on Bi-directional long short-term memory neural network and attention mechanism
Yu Guo, Dongfang Yang, Kun Zhao, Kai Wang
Abstract
At present, lithium-ion batteries (LIBs) play an irreplaceable role in various fields of production and life as an efficient energy storage element. The state of health (SOH) for LIB is critical to the safe operation of energy storage system. In fact, it is currently difficult to estimate SOH of LIB quickly and accurately. This paper proposes a method for SOH estimation that combines bidirectional long short-term memory (BiLSTM) neural network and attention mechanism. We extract three features from the incremental capacity (IC) curve as inputs to the model. The correlation rates between the proposed features and battery capacity are more than 0.98. Finally, the NASA dataset is introduced for experimental verification. The verification results demonstrate that the proposed method achieves accurate estimation of the SOH for LIBs. In the experimental results, the root mean square error (RMSE) and mean absolute percentage error (MAPE) of the proposed method can be as low as 0.0051 and 0.34%, respectively.