Configuration file

Let us consider the image below (this image and the real dimensions can be extracted from Fig. 52 in Benali2019 and Fig. 1c in Liu2022.

Grains and pore space configuration.

This image (2D) consists of 805x252 pixels, and the real dimensions (3D) are 6.74e-3 x 2.5e-3 x 0.03e-3 [m]. We remark that the image of the pattern used in the numerical simulations in Liu2022 has a much higher resolution.

The current implementation allows for the following input parameters:

"""Set the framework parameters"""
6.74e-3     #Image-related, length of the microsystem [m]
2.5e-3      #Image-related, height of the microsystem [m]
0.03e-3     #Image-related, depth of the microsystem [m]
160         #Image-related, threshold for converting the image to binary
1.0         #Image-related, rescaled factor for the input image
0           #Image-related, minimum size of the grain clusters
0           #Image-related, tolerance to approximate the border as polygon
0           #Image-related, tolerance to approximate the grains as polygon
1.0         #Figure-related, line width to show the contours in the produced figures
0.2e-3      #Device-related, width of the top and bottom channels in the micromodel device [m]
8e-6        #Mesh-related, mesh size [m]
1e-6        #Fluid-related, kinematic viscosity [Dynamic viscosity/fluid_density, m2/s]
1e-12       #Tracer-related, diffusion coefficient [m2/s]
5.0e-4      #Simulation-related, inlet boundary condition (Pressure/fluid_density, [Pa/(kg/m3)])
120.0       #Simulation-related, end time for the tracer simulation [s]
1.0         #Simulation-related, time interval to write the tracer results [s]
1e-7        #Solver-related, convergence criterium for the pressure solution in the numerical scheme for the Stokes simulation
1e-8        #Solver-related, convergence criterium for the velocity solution in the numerical scheme for the Stokes simulation
10000       #Solver-related, maximum number of iterations for the Stokes simulation in case the convergence criteria have not been reached
1.0         #Solver-related, time step in the numerical scheme for the Tracer simulation [s]

Warning

Do not remove # in each line (in the current implementation this is used for the reading of the parameters).

Image-related parameters

The three first parameters set the real dimensions of the microsystem. The next parameter sets the threshold value to convert the image to binary (valid values from 0 to 255). The rescaled factor parameter reduces the number of pixels of the image (valid values between 0 and 1, see skimage.transform.rescale. The minimum size of the grain clusters controls the number of pixels to consider for the internal grains (valid values are greater than 0, see porespy.filters.trim_small_clusters ). The two following parameters, for setting the tolerance to approximate the contours of polygons, reduce the number of points in the extracted border and internal grains respectively (valid values are 0 or greater than 0, see skimage.measure.approximate_polygon ). The following figure shows the internal grains and border for three decreasing values of the minimum size cluster and tolerances.

Extracted contours for minimum size of the cluster grains and tolerances for the polygon approximation of (top) 500, 5, 5, (middle) 100, 1, 1, and (bottom) 0, 0, 0, respectively.

Tip

The smaller are these numbers, the more detail is included in the mesh; however, this increases the execution time to create the mesh and to run the simulations.

Figure- and device-related parameters

The figure parameter controls the line width of the contours on the output images, while the device parameter controls the width of the channels in the device microsystem (see the following subsection for details about the implemented device).

Mesh-related parameters

Currently the only input parameter (via de configuration file, one can always modify the template files) for the mesh is the mesh size. The following figure shows the generated mesh for two different mesh sizes.

Two different generated meshes of sizes (top) 1e-4 and (bottom) 1e-5.

There are currently two template files that define the configuration of the microsystem: image.mako and device.mako. The template image.mako considers that the flow is from top to bottom of the image, while device.mako creates the micromodel geometry as in Benali2019 and Liu2022. On the grains and device walls we consider no-slip conditions. Then one could look at those files to modify the values (e.g., the flow direction or create quad elements for the mesh) or to define new micromodel geometries (e.g., a micromodel with a vertical cross-type shape). The following figure shows the geometry of the device micromodel.

Geometry of the device mode.

Remaining parameters

The remaining parameters are OpenFOAM related. Refer to the online OpenFOAM resources for details about the simulator and this nice presentation using the OpenFOAM solver simpleFoam in another micromodel application. Details about the solver simpleFoam and mathematical model can be found in this link. Details about the solver scalarTransportFoam and mathematical model can be found here.