Tools for Parameterization of a Single Particle Model with Electrolyte Dynamics (SPMe) with Bayesian Experimental Design to Capture Degradation-Induced Capacity Fade in Lithium-Ion Batteries

Abstract

The intelligent design of durable, next-generation lithium-ion batteries can be facilitated through the use of computationally efficient models that capture battery physics. In this work, we demonstrate a single particle model with electrolyte dynamics (SPMe), which accounts for battery capacity fade caused by mechanisms such as solid electrolyte interface (SEI) growth, lithium plating, and stress-induced particle cracking. The model is trained using both experimental data and data generated using a higher-fidelity Doyle-Fuller-Newman model so the ground truth parameters are known. To quantify and reduce the uncertainty in the unknown model parameters, we are developing a Bayesian experimental design approach that, with the aid of multiple models with varying fidelity, can rapidly identify optimal experimental input conditions leading to the greatest expected information gain.