Small-signal wide-band electrochemical-impedance measurements for battery management

Abstract

Batteries with lithium-iron-phosphate cathodes paired with lithium-titante anodes (LFP-LTO) produce a complex challenge for state-of-charge (SOC) estimation due to their flat open-circuit voltage. Research for the LFP-LTO chemistry shows that the electrochemical impedance spectra (EIS) is a possible tool for SOC estimation. The purpose of this thesis is to contribute three developments for the system identification and characterization of LFP-LTO cell chemistry, more specifically related to EIS. The first contribution is wide bandwidth EIS identification method that is developed and validated, using experimental and simulation techniques. The identification method consists of combining measurements of the voltage response to small-signal current perturbations at varying time-scales to obtain a wide-bandwidth EIS response. The second contribution is to experimentally evaluate the variation of the EIS with state-of-charge for LPF batteries that was shown using a physical model in [1], which also indicated dependence of the charge/discharge history. The history dependence is evaluated by extracting EIS for a cell at several states-of-charge with different charge/discharge profiles. The final contribution is a method for identifying and calibrating a state-of-charge prediction model for a LFP-LTO cell. This state-of-charge prediction model implements a LASSO regression to find the prediction mapping from EIS to state-of-charge.