A new gas–liquid dynamics model towards robust state of charge estimation of lithium-ion batteries

Abstract The accurate prediction of state of charge (SOC) is indispensable in the battery management system (BMS). Herein, a new gas–liquid dynamics (GLD) battery model based on the gas–liquid system is proposed to estimate SOC precisely and reliably for the lithium-ion battery (LIB). Electron transmission process, terminal voltage lag, lithium-ion diffusion and balance, as well as the ohmic resistance effect of LIBs can be clearly embodied in this model. Concurrently, the presented SOC estimator neither couples intelligent algorithms nor involves complicated matrix operations to guarantee the real-time performance of the online estimation. The genetic algorithm (GA) is adopted to identify model parameters. The estimation results of GLD model show the maximum errors of 1.74%, 3.02% and 2% under the Dynamic Stress Test (DST) cycle, the Urban Dynamometer Driving Schedule (UDDS) and constant current discharging test at 0.6–1.8C with the SOC reducing from 100% to 0, respectively. This model has the merits of simple structure, high-precision and strong robustness.