Cloud-based estimation of lithium-ion battery life for electric vehicles using equivalent circuit model and recurrent neural network

Abstract

With the increasing adoption of electric vehicles, the limitations of BMS in terms of storage capacity and computational power lead to a gradual accumulation of errors in battery capacity estimation over time. This growing inaccuracy significantly compromises the effective management of on-board power battery states. The challenge of battery capacity estimation based on extensive cloud-stored data has become a key focus in current research. In this paper, we propose an enhanced method that combines the ECM with a RNN to address this issue. By incorporating the relationship between OCV and SOC into the ECM, and employing PSO for direct capacity identification, we achieve accurate estimation. Furthermore, the use of RNN dynamically adjusts the observation noise in Kalman filtering, significantly improving the precision of the estimation. Experimental results demonstrate that the proposed `method yields a RMSE of <3 % and an average relative error below 2 %, compared to traditional approaches. This study presents a high-precision, efficient solution for estimating battery capacity using cloud-based data from electric vehicles, offering substantial application value.