A novel state-of-health prediction method based on long short-term memory network with attention mechanism for lithium-ion battery

The state-of-health (SOH) of lithium-ion batteries is one of the important core issues of battery management systems (BMS). After the battery reaches its end of life (EOL), its safety performance will deteriorate rapidly, which will be a huge threat to electric vehicles (EVs). Therefore, the accurate SOH prediction can ensure the safety and reliable operation of the battery, which is a critical and challenging issue. Accordingly, this paper proposes a novel SOH prediction method for lithium-ion batteries based on the long short-term memory (LSTM) neural network combined with attention mechanism (AM). First, moving average filter is applied to the lithium-ion battery capacity data for the purpose of reducing noise. Then, according to the battery capacity data of different datasets and different discharge rates, different weights are given to the LSTM hidden layer by AM to enhance the important information, so as to complete SOH prediction. Finally, the model is tested on new data and compared with the current data-driven prediction model. The experiment results show that the proposed SOH prediction method is more accurate, simple and robust. Furthermore, the SOH prediction method proposed in this paper is full of promising for practical EVs applications.