You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This notebook goes over most of the key aspects OpenPNM with minimal description, favoring links to other relevant notebooks and resources instead for further reading
Provides a explanation of how numerical data is stored in OpenPNM, including the storage of the topological information using sparse adjacency matrices
Labels are used to mark pores and throats for easier selection elsewhere in the simulation, such as applying boundary conditions. This tutorial illustrates how to use the labels that are automatically included on generated networks and how to add new user-defined labels.
One of the "power" features of OpenPNM is the ability to define multiple subdomains, enabling the simulation of layered or hierarchical materials with different pore sizes. This tutorial provides an explanation of this feature.
OpenPNM simulations called Projects and several Projects can be open within a single Workspace. This tutorial illustrates the relationship between these to constructs and their features.
The Cubic lattice is the classic pore network topology, and provides an excellent starting point for any investigation. The ability to alter the coordination number up to 26, combined with deleting random pores and throats creates an even more realistic topology.
The simultaneous simulation of transport in the void and solid, and their interactions, can be done with two interpenetrating cubic networks as illustrated in this tutorial
Cubic lattices do not have to be constrained to cubic shaped domains. This tutorial explains how arbitrary shaped domains (e.g. a spherical particle) can be created, though note the networks would still have a cubic lattice connectivity.
If truly random topology is necessary then this can be accomplished by performing a Delaunay tessellation on randomly (or not so randomly) distributed points
Boundary pores are useful when conducting transport simulations, but they are not added to generated OpenPNM network automatically. This tutorial explains their importance as well as how to add them to the network.
It's possible to alter a pore network by manually adding/removing pores and/or throats. This tutorial describes the available tools in OpenPNM and explains what occurs behind the scenes.
Given two or more networks, it is necessary to somehow join them into a single domain to perform simulations. This tutorial explains the difference between merging and stitching, and gives some examples on how to do both.
This example is one of the benchmark examples to illustrate the capability of OpenPNM for fast and accurate prediction of permeability. A network model for Benthiemer rock sample were extracted using PoreSpy. Permeability and of the network and effective diffusivity of air was then calculated implementing Stokes flow and Fickian diffusion in OpenPNM. Methods for calculation of permeability and diffusivity (effective properties) are included in generic transports in OpenPNM.
This example is a benchmark example to calculate the permeability of a Doddignton rock sample. The extracted network is loaded in OpenPNM for implementing Stokes flow.
This example is a benchmark example to calculate the permeability of a Berea rock sample. The extracted network is loaded in OpenPNM for implementing Stokes flow and Fickian diffusion for effective properties calculation.
This example is a benchmark example to illustrate the steps towards calculation of permeability for a Berea rock sample. This example includes the network extraction procedure in PoreSpy as well. OpenPNM was then used to find the permeability of the network.
This example illustrates the the steps for extracting networks from synthetic image of porous medium. Network extractions on digital porous media images can be implemented in PoreSpy open source package. PoreSpy output is consistent with OpenPNM input modules. The extracted network can then be easily used for other purposes in OpenPNM.
The basic SpheresAndCylinders class is a good place to start since it has all the necessary pore-scale models, but to customize it for a specific material requires adjusting the pore and throat size distribution. This example illustrates how to change the pore-size distribution function and how to tweak the distributions parameters to get a suitable result.
The Voronoi tessellation has been successfully used to model transport in fibrous media in several papers. This class generates a VoronoiDelaunay network object, then further creates a geometry object containing all the associated geometrical models.
OpenPNM includes only a few common phases as predefined classes, but it is quite easy to create new phases with custom properties. This example illustrates how to leverage the library of generic pore-scale models to create phases for any application.
Ordinary percolation corresponds to pressure-based invasion of non-wetting fluid. The OrdinaryPercolation class is the basis for simulating mercury intrusion and other capillary pressure curve measurements. This example illustrates its general use.
Invasion percolation corresponds to the to volume-based injection of a non-wetting phase. InvasionPercolation is the preferred approach for estimating multiphase transport properties since all saturations can be obtained, unlike OrdinaryPercolation which only produces saturations at discrete applied pressure steps.
Mixed invasion percolation is used for simulating imbibition of a wetting fluid. The mixed aspect refers to the fact that both pore bodies and throats are considered during the invasion.
The IntrusionPorosimetry class is a subclass of OrdinaryPercolation which is specifically developed for simulating capillary drainage of a non-wetting fluid. This class illustrates how to compare simulation to experimental data.
Calculating the rate of diffusion through a network is a primary use of OpenPNM, and this example illustrates how to do it, and how to convert the result into known transport parameters such as tortuosity and formation factor.
This example illustrates how to define a costum diffusivity model and assign to the phase object. A Fickian diffusion on a cubic network with a variable diffusivity is implemented.
The direct connection between the structure on the transport is one of main benefits of pore networks. This examples uses a 2-layered network with 3x smaller pores in on layer to illustrate how the permeability coefficient for each direction can be determined.
OpenPNM supports different multiphase transport phenomena. This example shows the procedures for calculation of relative diffusivity of Air and Water phases in a cubic network. For a multiphase diffusive mass transport problem invasion and distribution of phases can be modeled with percolation algorithms. Relative diffusivity of phases can then be calculated implementing Fickian diffusion for each phase considering multiphase pore scale models.
OpenPNM is capable of simulating multiphase transport mechanisms such as drainage and imbibition in porous media. Relative properties of phases can be calculated assigning a related pore scale model and transport algorithm. This example presents the calculation procedures for relative permeability of Air and Water phases for a 2D network. Calculations for 3D networks follow the same steps.
The package OpenPNM allows for the simulation of many transport phenomena in porous media such as Stokes flow, Fickian diffusion, advection-diffusion, transport of charged species, etc. Transient and steady-state simulations are both supported. An example of a transient Fickian diffusion simulation through a Cubic pore network is shown here.
OpenPNM supports adding reaction terms to both steady state and transient simulations. OpenPNM already includes many different source term models that can be added to simulate a reaction. In this example, we show how to add a powerlaw source term model to a transient fickian diffusion simulation.
Transient advection-diffusion simulations can be used to model elution curves. This example shows how to fit the concentration profile at the outlet to determine the dispersion coefficient of a network.
There are several functions built into OpenPNM for generating quick and dirty plots of network topology. These are useful for ensuring topological manipulations are working as expected, for example.
The Statoil format is used by the group at Imperial College, who have pioneered the principle of sharing code. Their maximal ball network extraction code outputs 4 files, which can be imported into OpenPNM, as outlined in this example.