Differential hysteresis models for a silicon-anode Li-ion battery cell

Voltage hysteresis in Li-ion battery cells makes it very difficult to estimate their state-of-charge (SOC) accurately. Hysteresis is prominent in next-generation chemistries, which have anodes with at least some silicon content. Good models can improve SOC-estimation accuracy, allowing longer use and enabling safer charge and discharge processes. This paper presents the first application of which we are aware of two computationally simple differential hysteresis models (vs. common operator-based models) to a Li//Si-graphite cell, and optimizes model parameter values in stages to get progressively better fits to measured data, ultimately arriving at SOC-dependent parameter values that fit the data well.