Prediction of Lithium-Ion Battery Capacity based on The ARMA Method

Lithium-ion battery capacity degradation is an inevitable phenomenon. How to predict lithium ion battery capacity more accurately is the main content of battery health management research. Current prediction methods usually require a lot of experimental data to have a good prediction accuracy, such as regression analysis and decomposition period. So how to make a more accurate battery remaining life prediction with fewer data is currently needed to study. This paper uses sample entropy and ARMA model for prediction. First, the sample entropy is used to analyze the discharge voltage curve, and the voltage change rule is analyzed. Second, the ARMA model is used to predict the short-term battery capacity change. Finally, the experimental results under different conditions are analyzed. By analyzing the experimental results, it is found that ARMA has higher accuracy for short-term forecasting based on fewer data.