DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Predicting Effective Diffusivity of Porous Media from Images by Deep Learning

Journal Article · · Scientific Reports
 [1];  [2]; ORCiD logo [3];  [4];  [1]
  1. Virginia Polytechnic Inst. and State Univ. (Virginia Tech), Blacksburg, VA (United States)
  2. Virginia Polytechnic Inst. and State Univ. (Virginia Tech), Blacksburg, VA (United States); Xi’an Jiaotong Univ., Xi’an (China)
  3. Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
  4. Xi’an Jiaotong Univ., Xi’an (China)

We report the application of machine learning methods for predicting the effective diffusivity (De) of two-dimensional porous media from images of their structures. Pore structures are built using reconstruction methods and represented as images, and their effective diffusivity is computed by lattice Boltzmann (LBM) simulations. The datasets thus generated are used to train convolutional neural network (CNN) models and evaluate their performance. The trained model predicts the effective diffusivity of porous structures with computational cost orders of magnitude lower than LBM simulations. The optimized model performs well on porous media with realistic topology, large variation of porosity (0.28–0.98), and effective diffusivity spanning more than one order of magnitude (0.1 ≲ De < 1), e.g., >95% of predicted De have truncated relative error of <10% when the true De is larger than 0.2. The CNN model provides better prediction than the empirical Bruggeman equation, especially for porous structure with small diffusivity. The relative error of CNN predictions, however, is rather high for structures with De < 0.1. To address this issue, the porosity of porous structures is encoded directly into the neural network but the performance is enhanced marginally. Further improvement, i.e., 70% of the CNN predictions for structures with true De < 0.1 have relative error <30%, is achieved by removing trapped regions and dead-end pathways using a simple algorithm. These results suggest that deep learning augmented by field knowledge can be a powerful technique for predicting the transport properties of porous media. In conclusion, directions for future research of machine learning in porous media are discussed based on detailed analysis of the performance of CNN models in the present work.

Research Organization:
Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
Sponsoring Organization:
USDOE Laboratory Directed Research and Development (LDRD) Program
Grant/Contract Number:
89233218CNA000001
OSTI ID:
1619666
Alternate ID(s):
OSTI ID: 1599031
Report Number(s):
LA-UR-19-23183
Journal Information:
Scientific Reports, Vol. 9, Issue 1; ISSN 2045-2322
Publisher:
Nature Publishing GroupCopyright Statement
Country of Publication:
United States
Language:
English

References (40)

Morphology, propagation dynamics and scaling characteristics of drying fronts in porous media: PROPAGATION DYNAMICS OF DRYING FRONTS journal May 2012
Seeing permeability from images: fast prediction with convolutional neural networks journal September 2018
Automatic Detection of Cerebral Microbleeds From MR Images via 3D Convolutional Neural Networks journal May 2016
Gas Diffusion in Porous Media journal July 1959
Mesoscopic predictions of the effective thermal conductivity for microscale random porous media journal March 2007
Pore Scale Simulation of Transport and Electrochemical Reactions in Reconstructed PEMFC Catalyst Layers journal January 2010
Deep learning journal May 2015
Fast connected-component labeling journal September 2009
Identity Mappings in Deep Residual Networks book January 2016
Permeability prediction of shale matrix reconstructed using the elementary building block model journal November 2015
X-ray imaging and analysis techniques for quantifying pore-scale structure and processes in subsurface porous medium systems journal January 2013
Improving direct physical properties prediction of heterogeneous materials from imaging data via convolutional neural network and a morphology-aware generative model journal July 2018
Gradient-based learning applied to document recognition journal January 1998
A multi-block lattice Boltzmann method for the thermal contact resistance at the interface of two solids journal June 2018
Deep learning approaches for mining structure-property linkages in high contrast composites from simulation datasets journal August 2018
Material structure-property linkages using three-dimensional convolutional neural networks journal March 2018
Pore Architecture and Connectivity in Gas Shale journal February 2015
Generalized lattice Boltzmann model for flow through tight porous media with Klinkenberg's effect journal March 2015
The permeability of fibrous porous media journal June 1986
Multi-column deep neural networks for image classification conference June 2012
Nanoscale simulation of shale transport properties using the lattice Boltzmann method: permeability and diffusivity journal January 2015
A dual-tube model for gas dynamics in fractured nanoporous shale formations journal October 2014
Fast Feature Pyramids for Object Detection journal August 2014
Gas Diffusion in Porous Media journal July 1959
Deep Learning text January 2018
The permeability of fibrous porous media journal June 1986
X-ray imaging and analysis techniques for quantifying pore-scale structure and processes in subsurface porous medium systems journal January 2013
A multi-block lattice Boltzmann method for the thermal contact resistance at the interface of two solids journal June 2018
Improving direct physical properties prediction of heterogeneous materials from imaging data via convolutional neural network and a morphology-aware generative model journal July 2018
Deep learning approaches for mining structure-property linkages in high contrast composites from simulation datasets journal August 2018
Permeability prediction of shale matrix reconstructed using the elementary building block model journal November 2015
Fast connected-component labeling journal September 2009
Seeing permeability from images: fast prediction with convolutional neural networks journal September 2018
A dual-tube model for gas dynamics in fractured nanoporous shale formations journal October 2014
Pore Architecture and Connectivity in Gas Shale journal February 2015
Gradient-based learning applied to document recognition journal January 1998
Gas Diffusion in Porous Media journal July 1959
Pore Scale Simulation of Transport and Electrochemical Reactions in Reconstructed PEMFC Catalyst Layers journal January 2010
Deep-learning methods for predicting permeability from 2D/3D binary-segmented images conference September 2016
Nanoscale simulation of shale transport properties using the lattice Boltzmann method: permeability and diffusivity text January 2014