UC Riverside

Lithium ion batteries are widely used in the world. The impedance of lithium ion batteries can potentially give informative insights of battery’s life and health if the interpretation of impedance is reliable enough. In this work, the impedance of Panasonic NCA 18650B battery was investigated through electrochemical impedance spectroscopy (EIS) and was modeled using artificial neural network. EIS study shows that under different cycling conditions, such as overcharge and overdischarge, the impedance change of electrolyte, electrode/electrolyte interface of the battery exhibit different behavior. Artificial neural network modeling gives an accurate prediction of battery future impedance for different sections such as equivalent series resistance (ESR), solid electrolyte interface (SEI) and charge transfer resistance, which can be used in battery state of health (SOH) estimation and prediction. Lastly, a design of impedance analysis circuit is given which make it possible to conduct real time impedance measurement on battery management systems (BMS).