Research on Lithium-Ion Battery State of Health Prediction Based on Autoregressive Integrated Moving Average Mode

Precise prediction of the state of health (SOH) of lithium-ion batteries exerts a key effect on ensuring the safe and reliable operation of battery systems. To solve the issue of low precision in traditional health forecasts for lithium-ion batteries, a lithium-ion battery health forecast approach based on the Autoregressive Integrated Moving Average (ARIMA) model is proposed. Firstly, the aging characteristics of batteries are analyzed based on the NASA and Maryland University public datasets. Secondly, the ARIMA model is validated using different aging data from two operating conditions to explore the coupling association between battery health and cycle count. According to experimental outcomes, the ARIMA prediction model realizes a precise forecast of the SOH of lithium-ion batteries. Compared to the Long Short-Term Memory (LSTM) neural network health prediction model, the CS2-37 battery prediction results show an improvement of 92.2% in Mean Absolute Error (MAE), 91.2% in Root Mean Square Error (RMSE), and 92.6% in Mean Absolute Percentage Error (MAPE).