Capacity Estimation of Lithium-Ion Batteries Based on an Optimal Voltage Section and LSTM Network

Abstract

Accurate capacity estimation is the cornerstone of attaining the state of health and remaining useful life of lithium-ion batteries. However, most of existing methods for battery capacity estimation are developed based on the fully charging/discharging condition, which is limited for onboard applications. This paper proposes a capacity estimation method based on an optimal voltage section. Firstly, the feasibility of capacity estimation based on sectional voltage data is demonstrated by correlation analysis between the voltage section-based health factors and the complete capacity. Secondly, the quantum particle swarm optimization algorithm is employed to determine the optimal voltage section. Thirdly, a mapping model between health factors and battery capacity is constructed using a long short-term memory neural network. Finally, validation results on public data sets show that the proposed method can realize accurate capacity estimation with an average root mean square error of 1.53%.