Poromics

Poromics estimates transport properties of 3D porous material images. It is GPU-accelerated and designed to be fast and easy to use.

Supported properties:

• Tortuosity / effective diffusivity — via Julia-based FD solver (tortuosity_fd) or Taichi-based LBM D3Q7 BGK solver (tortuosity_lbm)
• Absolute permeability — via Taichi-based LBM D3Q19 MRT solver (permeability_lbm)

Installation

The Julia-based FD solver depends on Tortuosity.jl, which is installed automatically. The LBM solvers use Taichi with automatic GPU detection.

Uv

Install uv, and then run the following command in a terminal/command prompt:

uv pip install poromics

Pip

If you prefer to use pip, run the following command in a terminal/command prompt:

pip install poromics

Basic Usage

Tortuosity (Julia FD solver)

import porespy as ps
import poromics

im = ps.generators.blobs(shape=[100, 100, 100], porosity=0.6)
result = poromics.tortuosity_fd(im, axis=0, rtol=1e-5, gpu=True)
print(result.tau, result.D_eff)

Tortuosity (LBM solver)

result = poromics.tortuosity_lbm(im, axis=0, D=1e-9, voxel_size=1e-6)
print(result.tau, result.D_eff)

Permeability (LBM solver)

result = poromics.permeability_lbm(im, axis=0, nu=1e-6, voxel_size=1e-6)
print(result.k)

Result objects

TortuosityResult attributes: im, axis, porosity, tau, D_eff, c, formation_factor, D.

PermeabilityResult attributes: im, axis, porosity, k, u_darcy, u_pore, velocity, pressure.

Simulation solvers

For more control, use the solver classes directly:

from poromics.simulation import TransientDiffusion, TransientFlow

solver = TransientDiffusion(im, axis=0, D=1e-9, voxel_size=1e-6)
solver.run(n_steps=100_000, tol=1e-2)
print(solver.concentration.shape, solver.converged)

CLI

poromics --help

Acknowledgments

The LBM solvers are based on taichi_LBM3D by Yi-Jie Huang.

Roadmap

• Diffusional tortuosity
  • Julia FD solver (Tortuosity.jl)
  • Taichi LBM D3Q7 BGK solver
• Transient tortuosity
  • Julia FD solver (Tortuosity.jl)
• Permeability
  • Taichi LBM D3Q19 MRT solver
• Electrode tortuosity
• Julia/Taichi coexistence via subprocess isolation
• Add command-line interface (CLI) for easy usage
• Add support for sysimage creation upon installation for faster startup