Lithium-ion Battery Remaining Useful Life Prediction Based on an Integrated Method

Abstract

As the main power source in many electronic and electrical devices, lithium-ion battery plays a critical role in thesafety and reliability of industrial systems. So it is of great importance to evaluate the performance degradation and predict the remaining useful life (RUL) for batteries. This paper developed a novelintegrated methodto realize RUL prediction by combining unscented Kalman filter (UKF), empirical mode decomposition (EMD) and relevance vector regression (RVR). First of all, UKF is used toestimate and adjust the system states andmaketheoriginal error residualseries. Then, the error series will be decomposedby EMD, and the decomposition result will be analyzedto produce a new error sequence whichwill beused by RVR to make prediction of the prognosticerror residual. Finally, UKF will adopt the predicted errorresidualto estimate the battery parametersrecursivelyand predict RUL. The result of theexperiment on a lithium battery shows that, the proposed method has excellent performance of reliability and prediction accuracy.