Deep learning for diffusion in porous media

Article Physics, Multidisciplinary

How Stickiness Can Speed Up Diffusion in Confined Systems

A. Alexandre et al.

Summary: Using a general adsorption desorption model with surface diffusion, this study analytically shows that making surfaces or obstacles attractive can accelerate dispersion.

PHYSICAL REVIEW LETTERS (2022)

Add to Collection

Article Physics, Applied

Convolutional neural network based prediction of effective diffusivity from microscope images

Smruti Ranjan Sethi et al.

Summary: This study introduces a method using deep learning to predict the effective diffusivity of porous media from a limited set of scanning electron microscope images. A convolutional neural network (CNN) model is trained and the images are classified and processed to improve the prediction accuracy.

JOURNAL OF APPLIED PHYSICS (2022)

Add to Collection

Article Multidisciplinary Sciences

Self-normalized density map (SNDM) for counting microbiological objects

Krzysztof M. Graczyk et al.

Summary: This study focuses on the statistical properties of the density map approach for counting microbiological objects on images. An improved self-normalized density map model is proposed, which can accurately predict the total number of objects in the image. The improved model outperforms the original model in terms of efficiency and statistical consistency.

SCIENTIFIC REPORTS (2022)

Add to Collection

Review Biochemistry & Molecular Biology

Transport in the Brain Extracellular Space: Diffusion, but Which Kind?

Eugene B. Postnikov et al.

Summary: The discussion of substance transport mechanisms in the brain parenchyma in recent years has mainly focused on the glymphatic concept and the role of diffusion, but there has been limited attention given to the potential variations in diffusion characteristics. New theoretical approaches have expanded the understanding in this field.

INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES (2022)

Add to Collection

Article Physics, Multidisciplinary

Collisions Enhance Self-Diffusion in Odd-Diffusive Systems

Erik Kalz et al.

Summary: It has been discovered that collisions can enhance self-diffusion in odd-diffusive systems, which goes against the common belief. Through analytical predictions and Brownian dynamics simulations, this counterintuitive behavior in the odd-diffusive system has been explained.

PHYSICAL REVIEW LETTERS (2022)

Add to Collection

Article Engineering, Chemical

Homogenization-Informed Convolutional Neural Networks for Estimation of Li-ion Battery Effective Properties

Ross M. Weber et al.

Summary: This study proposes a method based on convolutional neural networks (CNN) and homogenization theory to predict the effective properties of lithium-ion battery electrodes. Compared to traditional methods, this method has advantages in both accuracy and computational efficiency.

TRANSPORT IN POROUS MEDIA (2022)

Add to Collection

Article Multidisciplinary Sciences

Single-molecule tracking of Nodal and Lefty in live zebrafish embryos supports hindered diffusion model

Timo Kuhn et al.

Summary: This study uses reflected light-sheet microscopy to observe the movement and binding behavior of the activator-inhibitor signaling pair Nodal and Lefty in live developing zebrafish embryos. They conclude that the diffusion coefficients of molecules are high in extracellular cavities, but molecules still accumulate in these cavities.

NATURE COMMUNICATIONS (2022)

Add to Collection

Article Multidisciplinary Sciences

Inverse design of anisotropic spinodoid materials with prescribed diffusivity

Magnus Roding et al.

Summary: In this study, the effective diffusivity of spinodal decomposition-like structures with tunable anisotropy is predicted using a convolutional neural network. By incorporating the predictions into an approximate Bayesian computation framework for inverse problems, computationally efficient design of microstructures with prescribed diffusivity in all three directions is achieved.

SCIENTIFIC REPORTS (2022)

Add to Collection

Article Engineering, Chemical

Tortuosity of hierarchical porous materials: Diffusion experiments and random walk simulations

Veronique Wernert et al.

Summary: Diffusion experiments, effective medium theory calculations, and random walk simulations were conducted on hierarchical porous materials. The tortuosity, defined as the ratio of fluid mean square displacements in the absence and presence of the porous medium, was proposed as the main result of the mesoscopic simulations. The study compared the calculated tortuosity with experimental data and effective equations, and found a maximum value for the apparent tortuosity versus probe size/pore size ratio.

CHEMICAL ENGINEERING SCIENCE (2022)

Add to Collection

Article Engineering, Chemical

Physics- and image-based prediction of fluid flow and transport in complex porous membranes and materials by deep learning

Serveh Kamrava et al.

Summary: This paper introduces a deep learning network to predict the flow properties of porous membranes based on their morphology. By training the network with high-resolution images and pressure/velocity distribution data, accurate predictions for the desired properties can be achieved.

JOURNAL OF MEMBRANE SCIENCE (2021)

Add to Collection

Article Multidisciplinary Sciences

Objective comparison of methods to decode anomalous diffusion

Gorka Munoz-Gil et al.

Summary: Deviations from Brownian motion leading to anomalous diffusion are commonly found in transport dynamics, but challenging to characterize. An open competition comparing different approaches for single trajectory analysis showed that machine learning methods outperform classical approaches.

NATURE COMMUNICATIONS (2021)

Add to Collection

Article Engineering, Chemical

Computationally Efficient Multiscale Neural Networks Applied to Fluid Flow in Complex 3D Porous Media

Javier E. Santos et al.

Summary: The permeability of complex porous materials is crucial to various engineering disciplines. Direct flow simulation provides accurate results but is computationally expensive, while semi-analytical models are limited in applicability. Data-driven machine learning approaches show promise in building general models but face challenges in scaling to large 3D domains.

TRANSPORT IN POROUS MEDIA (2021)

Add to Collection

Article Chemistry, Physical

Simulating fluid flow in complex porous materials by integrating the governing equations with deep-layered machines

Serveh Kamrava et al.

Summary: A machine learning approach incorporating mass conservation and the Navier-Stokes equations is introduced for accurately predicting flow properties of porous media with reduced computation time. The method shows high accuracy in predictions and can be applied to different porous media with reliable results.

NPJ COMPUTATIONAL MATERIALS (2021)

Add to Collection

Article Engineering, Chemical