Step By Step Process for Formulating a DFN Model for a Lithium Ion Full Cell Battery

Abstract

The Doyle Fuller Newman (DFN) model has commonly been used to simulate the behavior of lithium ion batteries. However, the ways that parameters for the model are obtained vary widely across the field. In extreme cases, parameters are determined from a minimal experimental data set that fails to consider the validity of the model for other experimental conditions. This can lead to parameters that only perform well for the specified experimental technique. A process is outlined for determining parameters for a DFN model of a lithium ion full cell battery that uses a wide variety of experimental techniques to ensure parameters perform well over a variety of experiments. A recommended order that the parameters should be determined in is also provided. A NCM-graphite full cell battery is used as an example for the outlined process. Experimental data from manufactured half cells for each electrode is used to determine OCV curves, conductivity, exchange current densities, capacitances, and diffusivities for graphite, 1 M LiFEPO4 EC/DEC (1:1) electrolyte, and NCM. Test procedures included GITT, current interruption, cycling, and EIS. The parameters determined for each electrode and the electrolyte from half-cell experiments are validated using experimental data from manufactured full cells. The final model parameters are compared to existing parameters in literature with key differences being highlighted as well as reasons that the differences may have occurred.