Kernel dies after the .run

[EvgenyUgolkov]

Dear Team, good day. Many thanks for your previous response, it helped me to advance a little bit further, but i stuck again. Could you help me?

My goal is to calculate the absolute permeability of the 3D Micro-CT image of sandstone, which i upload from the .TIF file (i attached this file in the .zip archive). The image has a size of 400x400x400 voxels, what seems to be not very large. I use the following code, and after .run my Jupyter Notebook kernel dies after 1 second.

Dear Team, could you tell me if i am doing anything wrong here? Also, maybe you have some recommendations how i can make this process better? Maybe i should add or optimize anything for better/faster results?

Here the code and the file in zip archive go

With respect, Evgeny

berea.zip

import torch
import openpnm as op
import porespy as ps
import numpy as np
import skimage.io as skio 
import matplotlib.pyplot as plt

# import 3D bimary image (0 and 255 values) into the tensor and normalize it to 0 and 1 values

Image_3D = torch.tensor(skio.imread('berea.tif', plugin="tifffile")).unsqueeze(0).float()/255
voxel_size = 5.345e-6 # 5.345 micro meters

# Transfer 0 and 1 into the PoreSpy convention where 0 = grains and 1 = pores. Check with porosity function 
# if everything is correct. Transfer the torch tensor into numpy array

Image_3D = 1 - Image_3D
print('Image_3D porosity', ps.metrics.porosity(Image_3D))
Image_3D = Image_3D[0].numpy()

# Extract the network with the PoreSpy and assign boundaries

snow_output = ps.networks.snow2(Image_3D, voxel_size=voxel_size)
snow_output = ps.networks.label_boundaries(snow_output.network, labels=[['left', 'right'], ['front', 'back'], ['top', 'bottom']])

# Put the extracted network into the OpenPNM

pn = op.io.network_from_porespy(snow_output)

# Determine the parameters of the network

pn['pore.diameter'] = pn['pore.equivalent_diameter']
pn['throat.diameter'] = pn['throat.inscribed_diameter']
pn.add_model_collection(op.models.collections.geometry.pyramids_and_cuboids)
pn.add_model(propname='throat.length', model=op.models.geometry.throat_length.cubes_and_cuboids)
pn.add_model(propname='throat.lens_volume', model=op.models.geometry.throat_volume.lens)
pn.add_model(propname='throat.total_volume', model=op.models.geometry.throat_volume.cuboid)
pn.add_model(propname='throat.volume', model=op.models.misc.difference, props=['throat.total_volume', 'throat.lens_volume'])

# Check the network for isolated and disconnected pores. Delete (trim) these pores

h = op.utils.check_network_health(pn)
op.topotools.trim(network=pn, pores=h['isolated_pores'])
h = op.utils.check_network_health(pn)
op.topotools.trim(network=pn, pores=h['disconnected_pores'])

# Determine the properties of the water and determine conductance

water = op.phase.Water(network=pn)
pn.add_model(propname='throat.hydraulic_size_factors', model=op.models.geometry.hydraulic_size_factors.pyramids_and_cuboids)
water.add_model(propname='throat.hydraulic_conductance',model=op.models.physics.hydraulic_conductance.hagen_poiseuille)

# Determine the Stocks Flow and boundary conditions

flow = op.algorithms.StokesFlow(network=pn, phase=water)
Pin, Pout = (101_325, 0)
flow.set_value_BC(pores=pn.pores('left'), values=Pin)
flow.set_value_BC(pores=pn.pores('right'), values=Pout)

# Check that before .run everything seems to work fine

print('Till this point no errors')

# Run the simulaiton
flow.run()

[jgostick]

Can you clarify how far the script gets before crashing? Break it into separate cells and see which is the problem?

[JahanzebTariq]

so is there a need to mention the equation as mentioned in the above coding or using the standard library is enough?

[jgostick]

It should not have any effect...it will just overwrite the model that is already there.

[EvgenyUgolkov]

@jgostick Professor, good day. I found the code with Berea Sandstone which you sent to JahanzebTariq 2 month ago, it helped. Many thanks for support!

[jgostick]

Can you remind me what the solution was?

[EvgenyUgolkov]

@jgostick it was not the solution for the dying kernel, but it was a useful example which helped me to understand the sequence of actions. Somehow by chance i noticed that if i first run the simple example of absolute permeability from the github, (here is it)

and after it run my code, then kernel does not die. However, if i don't run this simple example first, kernel dies on my example. I don't know why is it, but just use. Maybe you know?