Health Prognostics for Lithium-ion Battery Based on Hybrid Data-driven Method

Accurate State of health prediction of lithium-ion battery provides a guarantee for the safe driving of electric vehicles. However, battery aging is a long-term complex process accompanied by capacity self-regeneration phenomenon, which also brings challenges to accurately prediction of SOH. Therefore, a novel SOH prediction method based on mode decomposition and hybrid machine learning is proposed in this paper. The original capacity data is decomposed by complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to obtain the trend item and detail item, which indicates the long-term degradation trend and self-regeneration effect. After that, long short-term memory(LSTM) and support vector regression (SVR) are employed for prediction of trend item and detail item respectively. The final SOH prediction is obtained by accumulating the prediction results of the trend item and the detail item. The effectiveness of the proposed method is verified by 2 different datasets. The prediction error of the proposed method is under 2%, which is less than the compared methods. The prediction results of different dataset show good accuracy, which indicates that the proposed method has high robustness, good accuracy, and applicability.