Labeled Image Dataset of Generated Porous Electrode Microstructures and Calculated Transport Parameters for Neural Network Training

A dataset of 100,000 computer-generated images that represent Lithium-ion battery microstructures is provided in the 'images.zip' file. The binary 64x64 TIFF images show sample porous electrodes, with roughly 1-20 randomly generated and placed particles. Analysis performed demonstrates that the images reflect realistic electrodes by comparing porosity and specific surface area in addition to visual inspection. This dataset includes images with a variety of particle sizes, roughness, and orientation to encapsulate various electrode chemistries.

For each image, the effective diffusion and conductivity tensors are calculated by solving a partial differential equation (PDE) closure problem. These calculated values are provided as .txt files in the 'diffusion.zip' compressed folder. This dataset can be used to examine the relationship between microstructural geometry and effective transport parameters of the porous media.