A Variational Bayes Based State-of-Charge Estimation for Lithium-Ion Batteries Without Sensing Current

Abstract

State-of-charge (SOC) estimation of lithium-ion batteries in portable devices without sensing the current is considered in this study. Unlike the traditional approach of separate estimation of the SOC and current, we firstly reformulate the problem as state estimation for the nonlinear system with an unknown input which refers to the current in this study, then a novel variational Bayes-based unscented Kalman filter (VB-UKF) is proposed to simultaneously estimate the SOC and the current input for the nonlinear lithium-ion battery system. Verifications of the SOC estimation performance are made by the experiments under the pulsed-discharge profile and urban dynamometer driving schedule profile. Experimental results show that the proposed VB-UKF algorithm is superior to the unscented recursive three-step filter (URTSF) in terms of convergence rate and estimation accuracy of the SOC and current. And the SOC root mean square errors of VB-UKF are bounded within ±3% after convergence which indicates the feasibility and effectiveness of the proposed method.

Keywords

• Battery
• state-of-charge
• unknown current
• unscented Kalman filter
• variational Bayes