Modelling of Li-ion batteries dynamics using impedance spectroscopy and pulse fitting: EVs application

One of the most relevant tasks that must be carried out by a Battery Management System (BMS) is the diagnosis of the battery state. An important part of the algorithms used for determining the State of Charge (SOC) or the State of Health (SOH) requires a cell model to run. The most precise is the model used, the best is the estimation achieved by the algorithm. In this paper, two techniques for obtaining a model of the cell dynamics and calculating its parameters are analyzed: the time domain characterization and the frequency domain or impedance-based characterization. Their principal characteristics and some relevant considerations to take into account are explained, as well as the obtained results. The performance of both models is compared in terms of the voltage error and the requirements to use them. Finally, a combined methodology is proposed to overcome the problems which can appear when each technique is employed. The resultant model is validated at 25 °C in all SOC range using real measurements of a 40 Ah Li-ion cell with different current profiles, including pulses of diverse lengths and FUDS driving cycles. The tests show small error between the real response of the cell and the output of the model.