A cross-platform, high-performance SPH toolkit for image-based flow simulations on the pore scale of porous media

Osorno, Maria; Schirwon, Malte; Kijanski, Nadine; Sivanesapillai, Rakulan; Steeb, Holger; Göddeke, Dominik

doi:10.1016/j.cpc.2021.108059

Cited by 9 publications

(7 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The implementation includes the calculation of kernel, computation of field functions ( ρf , p, vf , v f ) and time integration whereas neighbour search and load balancing is handled by the HOOMDblue backend. A sufficient scaling behaviour as well as the numerical accuracy is demonstrated [4].…”

Section: Computational Aspectsmentioning

confidence: 90%

Simulation of weak‐inertia single‐phase flow in porous materials using Smoothed Particle Hydrodynamics

Krach

Steeb

2021

Proc Appl Math and Mech

View full text Add to dashboard Cite

Intrinsic permeability of a porous material is a crucial material parameter in various application fields like e.g. geosciences, materials science, biological sciences and mechanical engineering. In contrast to the permeability often used in purely linear models, however, it is also a (nonlinear) function of the Reynolds number (Re) with the onset of the influence of inertial forces. In coarse-grained continuum models the Forchheimer equation describes this incipient influence of viscous momentum exchange but it is based on the adaptation of experimentally obtained data and its range of validity remains a subject for discussion. Performing three-dimensional Direct Numerical Simulations for single-phase flow through porous media this work aims to show a smooth transition of effective intrinsic permeability within the interval 0.1 < Re < 1000. We use the fully-Lagrangian meshless Smoothed Particle Hydrodynamics method (SPH) to solve the weakly compressible Navier-Stokes equations for pore-scale resolved flow in the so-called weak-inertia regime. Our implementation utilizes HOOMD-blue thus allows us to perform massively parallel computations which are crucial for these applications.

show abstract

Section: Computational Aspectsmentioning

confidence: 90%

Simulation of weak‐inertia single‐phase flow in porous materials using Smoothed Particle Hydrodynamics

Krach

Steeb

2021

Proc Appl Math and Mech

View full text Add to dashboard Cite

show abstract

“…With the help of this discretization method and the derivation of the kernel function, divergence and gradient operators can likewise be discretized. We obtain (10) and see directly that the gradients of field functions do not have to be computed but merely the gradients of the kernel, which is an intrinsic advantage of the SPH method. By application of Eq.…”

Section: Smoothed Particle Hydrodynamics For Navier-stokes Equationsmentioning

confidence: 95%

“…For more information about the implementation and especially about the strong and weak scalability on CPU and GPU clusters for different geometries we refer to recent publications of the authors, cf. Schirwon [12] and Osorno et al [10].…”

Section: Sph Implementation In Hoomd-bluementioning

confidence: 99%

Comparing methods for permeability computation of porous materials and their limitations

Krach

Steeb

2023

Proc Appl Math and Mech

View full text Add to dashboard Cite

Efficient numerical simulations of fluid flow on the pore scale allow for the numerical estimation of effective material properties of porous media, e.g. intrinsic permeability or tortuosity. These parameters are essential for various applications where hydro-mechanical properties on larger scales have to be known. Numerical tools based intrinsically on pore scale simulations are known e.g. as Digital Rock Physics in geosciences and have even more and more replaced physical experiments. For these reasons, the validation of numerical methods as well as the establishment of clear limits regarding the application areas play an important role. Here, we compute single-phase flow through a porous matrix, e.g. irregular sphere packings, sandstones, artificially created thin porous media, on the pore scale. Therefore we implement on the one hand a Smoothed Particle Hydrodynamics algorithm for solving the Navier-Stokes equations and on the other hand a Finite Difference solver for the Stokes equations. Both methods work directly and seamlessly on voxel data of porous materials which are generated by µXRCT-scans or by microfluidic experiments that have undergone segmentation and binarization. We compare both solvers from a parallel performance point of view as well as their results for flows in the Darcy regime. In addition, we investigate the limitations of the solvers using the example of a porous material whose pore geometry changes over time and precipitation affects the flow conditions.

show abstract

“…For the simulation, the HOOSPH code was used in which the specific SPH-particle-particle interactions are implemented and that is based on HOOMD-blue (Anderson et al, 2008;Glaser et al, 2015). This is a rather computationally expensive method but has shown to be highly suitable for effective permeability computations of porous rocks in the Darcy and weak inertia regime as well as with onset of more pronounced inertia effects (Sivanesapillai et al, 2014;Osorno et al, 2020).…”

Section: Modelling Permeability Thermal and Effective Elastic Rock Propertiesmentioning

confidence: 99%

Digital Rock Physics: A Geological Driven Workflow for the Segmentation of Anisotropic Ruhr Sandstone

et al. 2021

View full text Add to dashboard Cite

Over the last 3 decades, Digital Rock Physics (DRP) has become a complementary part of the characterization of reservoir rocks due to the non-destructive testing character of this technique. The use of high-resolution X-ray Computed Tomography (XRCT) has become widely accepted to create a digital twin of the material under investigation. Compared to other imaging techniques, XRCT technology allows a location-dependent resolution of the individual material particles in volume. However, there are still challenges in assigning physical properties to a particular voxel within the digital twin, due to standard histogram analysis or sub-resolution features in the rock. For this reason, high-resolution image-based data from XRCT, transmitted-light microscope, Scanning Electron Microscope (SEM) as well as geological input properties like geological diagenesis, mineralogical composition, sample’s microfabrics, and estimated sample’s porosity are combined to obtain an optimal spatial segmented image of the studied Ruhr sandstone. Based on a homogeneity test, which corresponds to the evaluation of the gray-scale image histogram, the preferred scan sample sizes in terms of permeability, thermal, and effective elastic rock properties are determined. In addition, these numerically derived property predictions are compared with laboratory measurements to obtain possible upper limits for sample size, segmentation accuracy, and a geometrically calibrated digital twin of the Ruhr sandstone. The comparison corresponding gray-scale image histograms as a function of sample sizes with the corresponding advanced numerical simulations provides a unique workflow for reservoir characterization of the Ruhr sandstone.

show abstract

A cross-platform, high-performance SPH toolkit for image-based flow simulations on the pore scale of porous media

Cited by 9 publications

References 33 publications

Simulation of weak‐inertia single‐phase flow in porous materials using Smoothed Particle Hydrodynamics

Simulation of weak‐inertia single‐phase flow in porous materials using Smoothed Particle Hydrodynamics

Comparing methods for permeability computation of porous materials and their limitations

Digital Rock Physics: A Geological Driven Workflow for the Segmentation of Anisotropic Ruhr Sandstone

Contact Info

Product

Resources

About