Modeling a lithium-ion battery based on a threshold model

Modeling a lithium-ion (Li-ion) battery is the core issue for electric vehicles applications. The valid battery model and accurate model's parameters can improve accuracy of the state of charge (SOC) estimation and thus promote the commercialization of electric vehicles. Considering an electrical battery model with two resistance-capacitance (RC) parallel networks, the series resistance and the open-circuit voltage (OCV)-SOC function, the accurate estimate of battery parameters through a threshold model, namely threshold autoregressive and moving average with exogenous inputs (TARMAX) model is proposed, in this paper, which only needs online terminal voltage and current data. The approach is built by dividing voltage data into several regimes according to the thresholds. For each regime, a linear ARMA model is established which can be represented by the state-space equation for the linear battery model. Thus, the battery's parameters can be identified. To facilitate obtaining the time-variant model parameters, the terminal voltage and current data are collected by the sliding window. Finally, the effectiveness of the proposed modeling approach is verified by the simulation.