Battery management system in the Bayesian paradigm: Part I: SOC estimation

Accurate State-of-Charge (SOC) estimation of Li-ion batteries is essential for effective battery control and energy management of electric and hybrid electric vehicles. To this end, first, the battery is modelled by an OCV-R-RC equivalent circuit. Then, a dual Bayesian estimation scheme is developed-The battery model parameters are identified online and fed to the SOC estimator, the output of which is then fed back to the parameter identifier. Both parameter identification and SOC estimation are treated in a Bayesian framework. The square-root recursive least-squares estimator and the extended Kalman-Bucy filter are systematically paired up for the first time in the battery management literature to tackle the SOC estimation problem. The proposed method is finally compared with the convectional Coulomb counting method. The results indicate that the proposed method significantly outperforms the Coulomb counting method in terms of accuracy and robustness.