A Cloud-based State of Health Estimation Method for Lithium-ion Batteries Using Sparse Charging Data

Abstract

Under the current background of the national standard requirements for electric vehicle remote management and the companies’ needs for their digital platform services, automobile factories habitually transmit battery data to a cloud in a sparse way for state of health (SOH) estimation. Compared with the dynamic discharge condition of the vehicle, the charging process has been noticed in recent years due to its relatively stable advantages. In this paper, a cloud-based SOH estimation method for lithium-ion batteries using sparse charging data is proposed. A HI feature that can be extracted from the sparse data is derived and only the constant voltage charging process is considered. By using the experimental data deployed on the cloud computing system, the correlation of battery health with the HI is verified and SOH can be efficiently estimated by taking as this feature an input to three data-driven methods. The results show that all methods can achieve a low test error of 2% under 10s sparsity, which is promising to be applied in cloud sparse charging conditions.