Reducing the coordination number below Z=2.0

[LeonardoPEQ]

Recently, I’ve been working on a problem where I need to perform absolute permeability simulations on simple cubic networks ranging from 10x10x10 to 70x70x70. In these simulations, I need to compute the permeability coefficient for a variety of mean coordination numbers, including Z = 6, 5, 4, 3, 2, 1.75, 1.65, and 1.5.

However, I’ve run into an issue: the method suggested in the documentation doesn’t seem to work when the mean coordination number drops below 2. I suspect this is due to how the code constructs the network. Since it ensures all pores are connected by creating a minimum spanning tree, the network can’t achieve a mean coordination number lower than 2.

I’m wondering if you have any suggestions for how I might approach this differently. Alternatively, are there any plans to support networks with mean coordination numbers below 2 in future releases?

Here’s what I’ve tried so far in an attempt to reduce the coordination number. When Z < 2 (specifically, for 1.75, 1.65, and 1.5), the method seems to stop at Z = 2 and won’t go any lower:

import openpnm as op
import numpy as np

N, spacing, connectivity = 10, 1e-4, 6
pn = op.network.Cubic(shape=[N, N, N], spacing=spacing, connectivity=connectivity)
pn.add_model_collection(op.models.collections.geometry.spheres_and_cylinders)
pn.regenerate_models()

drop = op.topotools.reduce_coordination(pn, 1.7)
op.topotools.trim(network=pn, throats=drop)
pn.regenerate_models()

N_mean_all = np.mean(pn.num_neighbors(pn.Ps, mode='or', flatten=False))
print(N_mean_all)

In this case, the output is 1.998.

Any insights or advice would be greatly appreciated!

[jgostick]

The reason we use the spanning tree is so that the net work stays fully connected and hence viable for simulations. If you can find a different way to identify throats for deletion then that is an option. What I mean is that there is nothing special about the mst. For instance, you could select pores at random, then delete all their throat but 1 or 2. Repeat X times...THEN do the mst coordination number trick to bring the total coordination down. Perhaps there is already and algorithm for finding spanning trees other than the minimum one?