3D image-based simulations of battery electrodes have shown immense utility in providing insight into the reality of as-manufactured cells. However, transforming 3D images into computable meshes is an inherently uncertain process.
In our new paper (https://authors.elsevier.com/c/1idHR8Z1RY5j8F, free for a few weeks) in a special issue of Energy Storage Materials focused on machine learning, we employ multiple traditional computer vision and machine learning techniques to assess this uncertainty for graphite electrodes.
We combine this uncertainty with the intrinsic heterogeneity of a coated electrode to assess overall impacts to cell performance.