A cloud capacity estimation method for electric vehicle lithium-ion battery independent of cloud SOC

Lithium-ion batteries are the primary energy source for electric vehicles (EVs), and the available capacity estimation of each battery cell from power battery modules plays a vital role in battery management and lifespan prediction. Concerning conventional capacity estimation methods, including the ampere-hour integral method which uses cloud current and state of charge (SOC) data, the estimation error depends on the SOC accuracy. Moreover, in estimation methods based on traditional equivalent circuit models, the accuracy and stability decrease significantly over a wide temperature range. In this paper, the lithium-ion battery parameters including capacity and SOC, are estimated directly by a model combined with an optimization algorithm, employing cloud data of EVs. Firstly, an improved thermal-electric coupling equivalent circuit model (TEECM) based on conventional equivalent circuit model has been established, which improves the MAE for simulation accuracy by a maximum of 58 % in a wide temperature range. Additionally, considering the fluctuations to a certain extent in the results of capacity estimation under different conditions, the characteristics and estimation errors of data segments under different charging conditions are analyzed based on experimental data. The fuzzy logic algorithm is used to evaluate the weight of each data segment estimation, which is the observation noise of the Kalman filtering algorithm, the battery capacity estimated by the cloud data of EVs is modified by weighted filtering, effectively improving the accuracy and stability of lithium-ion batteries capacity estimation. Finally, the cloud data of a certain operating EV is used for capacity estimation, while conventional methods fail to converge, the proposed method ensures both greater stability and accuracy.