Research on Life Prediction of Lithium-ion Battery based on WEMD-ARIMA Model

Abstract

With the development of AGV and AMR, lithium-ion battery, as the main energy source of mobile robot, has also received extensive attention. In view of the low accuracy of the current prediction methods for the remaining cycle life of lithium-ion batteries, a new prediction model for the remaining cycle life of lithium-ion batteries is proposed in this paper. The cycle life degradation data of lithium-ion battery is regarded as a set of random time series. Firstly, the time series are decomposed by WEMD; They are decomposed into several IMF subsequences. The WEMD model proposed in this paper is improvement by adding white noise to the sequence on the basis of EMD. It makes the decomposed IMF sequence without modal mixing. Then, use the ARIMA prediction to predict each IMF subsequence; Finally, the prediction results of IMFs are superimposed to obtain the final prediction results. In this paper, experiments are carried out using NASA open cycle life degradation data of lithium-ion battery. The experimental results show that the model has better accuracy. The MAE and MAPE of B0005 and B0006 batteries are almost less than 0.01 and 1.