A hybrid model based on support vector regression and differential evolution for remaining useful lifetime prediction of lithium-ion batteries

Remaining useful life prediction plays an important role in battery management system. The fusion prognostics method has become a main research direction for improving the prediction performance. We present a hybrid model based on support vector regression and differential evolution to predict the remaining useful life of Li-ion battery, where differential evolution algorithm is used to obtain the support vector regression kernel parameters. The capacity, voltage, and current on discharge operation are considered in this study. Three Li-ion batteries from NASA Ames Prognostics Center of Excellence are used to illustrate the application. The results show that the proposed method has better prediction accuracy than the ten published methods. Regeneration factor has insignificant influence on the prediction accuracy of the proposed hybrid model.