Experimental Validation

This section collects benchmarks against experimental data rather than against other software.

In voids, validation asks a different question from software verification:

• software verification checks whether the implementation is numerically consistent with a reference workflow
• experimental validation checks whether the full image-to-network workflow predicts measured porosity or permeability closely enough to be scientifically useful

That means a validation mismatch is not automatically a software bug. It can come from any part of the workflow, including:

• grayscale preprocessing and segmentation assumptions
• ROI selection and representativeness
• network extraction topology
• pore/throat geometry assignment
• hydraulic conductance closure
• the reduction from a voxel image to a pore-throat graph

Current validation studies:

• DRP-317 sandstone validation overview
• DRP-317 Parker notebook report
• DRP-317 Kirby notebook report
• DRP-317 Bandera Brown notebook report
• DRP-317 Berea Sister Gray notebook report
• DRP-317 Berea Upper Gray notebook report
• DRP-317 Berea notebook report
• DRP-317 Castlegate notebook report
• DRP-317 Buff Berea notebook report
• DRP-317 Leopard notebook report
• DRP-317 Bentheimer notebook report
• DRP-317 Bandera Gray notebook report

The DRP-317 pages use these cited sources:

• Dataset: Neumann, R., ANDREETA, M., Lucas-Oliveira, E. (2020, October 7). 11 Sandstones: raw, filtered and segmented data [Dataset]. Digital Porous Media Portal. https://www.doi.org/10.17612/f4h1-w124
• Experimental reference paper: Neumann, R. F., Barsi-Andreeta, M., Lucas-Oliveira, E., Barbalho, H., Trevizan, W. A., Bonagamba, T. J., & Steiner, M. B. (2021). High accuracy capillary network representation in digital rock reveals permeability scaling functions. Scientific Reports, 11, 11370. https://doi.org/10.1038/s41598-021-90090-0