A novel streamlined particle‐unscented Kalman filtering method for the available energy prediction of lithium‐ion batteries considering the time‐varying temperature‐current influence

Abstract

Effective energy prediction is of great importance for the operational status monitoring of high-power lithium-ion battery packs. It should be embedded in the battery system performance evaluation, energy management, and safety protection. A new Streamlined Particle-Unscented Kalman Filtering method is proposed to predict the available energy of lithium-ion batteries, in which an Adaptive-Dual Unscented Transform treatment is conducted to realize the precise mathematical expression of its working conditions. For the accurate mathematical description purpose, an improved Synthetic-Electrical Equivalent Circuit modeling method is introduced into the internal effect equivalent process considering the influence of time-varying temperature and current conditions. As can be known from the experimental results, the proposed prediction method has a maximum estimation error of 2.27% and an average error of 0.80%, for the complex varying-current Beijing Bus Dynamic Stress Test. Under the Urban Dynamometer Driving Schedule working conditions, the available energy prediction has high accuracy with a maximum error of 1.83% and a voltage traction error of 3.28%. It provides vehicle-mounted available energy prediction schemes for effective management and safety protection of high-power lithium-ion batteries. Highlights A new Streamlined Particle-Unscented Kalman Filtering method is proposed to predict the available energy of lithium-ion batteries. Improved Synthetic-Electrical Equivalent Circuit modeling strategies are established to describe the nonlinear battery characteristics. Adopted predictive correction is investigated by considering the time-varying temperature and current influence. For effective convergence, an adaptive windowing function factor is introduced into the correction process with a maximum estimation error of 2.27% and an average error of 0.80% for the complex varying-current Beijing Bus Dynamic Stress Test working conditions. The vehicle battery available energy prediction is realized with a maximum error of 1.83% and a maximum voltage traction error of 3.28% for the Urban Dynamometer Driving Schedule working conditions.