Abstract: To improve the online capacity estimation accuracy of lithium-ion batteries, an adaptive capacity estimation scheme combining the online capacity identification based on the features of a few charging curves with Arrhenius capacity decay model is proposed. In view of the seldom situations of complete charging in battery electric vehicles, an online capacity identification scheme based on voltage characteristic points of constant-current charging curves is put forward. The scheme uses genetic algorithm to optimize the voltage characteristic points of the scaled and translated charging curves first and then online identify the present capacity of the battery by monitoring the constant-current charging data regarding the two fixed voltage characteristic points. For further enhancing the accuracy of online capacity estimation, an incremental PID algorithm is used to fuse the online capacity estimation and Arrhenius model to perform the closed-loop correction of model parameters. Finally, the results of cycle life experiment under alternating temperature condition show that the maximum estimation error of the proposed adaptive estimation scheme is less than 2%.

Key words: capacity estimation, cycle life, Arrhenius model, model parameters, incremental PID