A novel fusion prognostic approach for the prediction of the remaining useful life of a lithiumion battery

Abstract

The lithium-ion battery has been widely used in electronic devices. Remaining useful life (RUL) prediction allows for predictive maintenance of electronic devices, thus reducing expensive unscheduled maintenance. RUL prediction of the lithium-ion battery appears to be a hot issue attracting more and more attention as well as being of great challenge. In this paper, a new fusion prognostic approach based on error-correction is proposed to predict the RUL of lithium-ion battery, which combines unscented Kalman filter (UKF) with BP neural network. Firstly, UKF algorithm is employed to obtain prognosis based on an estimated model and build a raw error series. Next, the error series is utilized by BP neural network to predict the UKF future residuals, which remain zero without consideration. Finally, the prognostic residual is adopted to correct the prognostic result achieved by UKF. According to the remaining useful life prediction experiments for batteries, the fusion method has high reliability and prediction accuracy.