Remaining Useful Life Prediction for Lithium-Ion Batteries Based on Empirical Model and Improved Least Squares Support Vector Machine

Abstract

Remaining useful life (RUL) prediction for lithium-ion batteries is important for the safety of the battery. To improve the accuracy of multi-step prediction, a novel method based on empirical model and improved least squares-support vector machine (LS-SVM) is proposed. Firstly, we constructed the deviation capacity based on the third-order polynomial empirical model to reduce the fluctuation of the data. Secondly, the differential evolution (DE) algorithm and multi-kernel (MK) functions were used to optimize the LS-SVM, and the improved LS-SVM was then built, which improved the prediction accuracy of the LS-SVM. On this basis, the framework for RUL prediction was constructed successfully. Finally, the experimental verification was carried out. The results show that the prediction accuracy of the proposed method is better than that of the existing methods.KeywordsLithium-ion batteryRemaining useful life predictionMulti-kernel least squares support vector machineDifferential evolution algorithm