Online Model-Based Lithium-Ion Battery State-of-Charge Estimation Using RLS With Adaptive Three Forgetting Schemes

Le résumé fourni par la source

Accurate state-of-charge (soc) estimation of lithium-ion batteries is crucial, and whether the battery model parameters can be accurately identified will directly affect the results of soc estimation. In the field of soc estimation, recursive least squares (RLS) algorithm with a single forgetting factor is commonly used for battery parameter identification, which relies on defaulting to the same change rate for each battery parameter, whereas the change rate of each parameter is different during actual operation. After the model conversion in this paper, the input-output mapping contains three parameters with different change rates that need to be identified, therefore, a new method for parameter identification by three adaptive forgetting factors RLS is proposed, matching the three forgetting factors for three parameters with different change rates. To cope with the changing characteristics of the parameters, an adaptive strategy is proposed for adjusting values of forgetting factors, based on updating according to estimation error, the effects of influencing factors like soc are extended into the adaptive framework in the way of a function. Convergence analysis is provided, which can prove that the proposed algorithm can converge asymptotically with bounded error. Accurate soc estimation is further achieved using the noise adaptive Kalman filter algorithm. To verify the effectiveness of proposed method, experiments are conducted under different operating conditions, temperatures and aged batteries, whose results show that the proposed method can estimate the soc with an absolute error of less than 0.2%.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.