Bayesian Parameterization of Continuum Battery Models from Featurized Electrochemical Measurements Considering Noise**

Abstract

Physico-chemical continuum battery models are typically parameterized by manual fits, relying on the individual expertise of researchers. In this article, we introduce a computer algorithm that directly utilizes the experience of battery researchers to extract information from experimental data reproducibly. We extend Bayesian Optimization (BOLFI) with Expectation Propagation (EP) to create a black-box optimizer suited for modular continuum battery models. Standard approaches compare the experimental data in its raw entirety to the model simulations.By dividing the data into physics-based features, our datadriven approach uses orders of magnitude less simulations. For validation, we process full-cell GITT measurements to characterize the diffusivities of both electrodes non-destructively. Our algorithm enables experimentators and theoreticians to investigate, verify, and record their insights. We intend this algorithm to be a tool for the accessible evaluation of experimental databases.

Multiscale Investigation of Sodium‐Ion Battery Anodes: Analytical Techniques and Applications

“…The recent rise of machine learning techniques has enabled the development of highly accurate continuum models by improving parametrization processes and experimental correlation, and will allow for more widespread application in the near future. [ 331 ] This is the coarsest‐grained and most computationally inexpensive method that we will discuss, it can provide spatially‐ and time‐dependent system properties at larger scales than any of the previously mentioned methods. Continuum models are often used in conjunction with experiments, which allows for the identification of battery properties, such as SEI thickness, and processes, such as reaction kinetics.…”

How machine learning can extend electroanalytical measurements beyond analytical interpretation

“…In recent years, a few studies 24,34 have employed Bayesian optimization to estimate material properties (consistent with physics-based governing equations) of battery materials from experiments. Given that such an approach is still not widely used, we first illustrate essential concepts using a simple example.…”

Efficient Workflows for Detecting Li Depositions in Lithium-Ion Batteries

“…Additionally, datadriven techniques can be used to inform and enhance the models. 193,232 In case of commercial cells (as shown in Fig. 11), the effort is higher, since cell disassembly is needed to obtain the material composition in the cells.…”