scholarly journals Reconstruction of Three-Dimensional Porous Media Using Deep Transfer Learning

Geofluids ◽  
2020 ◽  
Vol 2020 ◽  
pp. 1-22
Author(s):  
Yi Du ◽  
Jie Chen ◽  
Ting Zhang
Keyword(s):  
Porous Media ◽  
Transfer Learning ◽  
High Efficiency ◽  
Three Dimensional ◽  
Fluid Flows ◽  
Training Image ◽  
Multiple Point ◽  
Reconstruction Methods ◽  
Engineering Sciences ◽  
Complex Features

The reconstruction of porous media is widely used in the study of fluid flows and engineering sciences. Some traditional reconstruction methods for porous media use the features extracted from real natural porous media and copy them to realize reconstructions. Currently, as one of the important branches of machine learning methods, the deep transfer learning (DTL) method has shown good performance in extracting features and transferring them to the predicted objects, which can be used for the reconstruction of porous media. Hence, a method for reconstructing porous media is presented by applying DTL to extract features from a training image (TI) of porous media to replace the process of scanning a TI for different patterns as in multiple-point statistical methods. The deep neural network is practically used to extract the complex features from the TI of porous media, and then, a reconstructed result can be obtained by transfer learning through copying these features. The proposed method was evaluated on shale and sandstone samples by comparing multiple-point connectivity functions, variogram curves, permeability, porosity, etc. The experimental results show that the proposed method is of high efficiency while preserving similar features with the target image, shortening reconstruction time, and reducing the burdens on CPU.

The MPS Reconstruction of Porous Media Using Multiple-Grid Templates

2013 ◽  
Vol 462-463 ◽  
pp. 462-465 ◽  
Author(s):  
Yi Du ◽  
Ting Zhang
Keyword(s):  
Porous Media ◽  
Large Scale ◽  
Grid Method ◽  
Training Image ◽  
Multiple Point ◽  
Grid Data ◽  
Large Scale Structures ◽  
Complex Features ◽  
Grid Nodes ◽  
Data Template

It is difficult to reconstruct the unknown information only by some sparse known data in the reconstruction of porous media. Multiple-point geostatistics (MPS) has been proved to be a powerful tool to capture curvilinear structures or complex features in training images. One solution to capture large-scale structures while considering a data template with a reasonably small number of grid nodes is provided by the multiple-grid method. This method consists in scanning a training image using increasingly finer multiple-grid data templates instead of a big and dense data template. The experimental results demonstrate that multiple-grid data templates and MPS are practical in porous media reconstruction.

Reconstruction of 3D Porous Media Using Multiple-Point Statistics Based on a 2D Training Image

2018 ◽  
Author(s):  
Yuqi Wu ◽  
Chengyan Lin ◽  
Lihua Ren ◽  
Weichao Tian ◽  
Yang Wang ◽  
...  
Keyword(s):  
Porous Media ◽  
Training Image ◽  
Multiple Point ◽  
Multiple Point Statistics

Reconstruction of 3D porous media using multiple-point statistics based on a 3D training image

2018 ◽  
Vol 51 ◽  
pp. 129-140 ◽  
Author(s):  
Yuqi Wu ◽  
Chengyan Lin ◽  
Lihua Ren ◽  
Weichao Yan ◽  
Senyou An ◽  
...  
Keyword(s):  
Porous Media ◽  
Training Image ◽  
Multiple Point ◽  
Multiple Point Statistics

scholarly journals Mesoscale Anisotropy in Porous Media Made of Clay Minerals. A Numerical Study Constrained by Experimental Data

Materials ◽  
2018 ◽  
Vol 11 (10) ◽  
pp. 1972 ◽  
Author(s):  
Thomas Dabat ◽  
Arnaud Mazurier ◽  
Fabien Hubert ◽  
Emmanuel Tertre ◽  
Brian Grégoire ◽  
...  
Keyword(s):  
Porous Media ◽  
Clay Minerals ◽  
Numerical Study ◽  
Three Dimensional ◽  
Particle Orientation ◽  
Anisotropic Properties ◽  
Engineering Sciences ◽  
Particle Anisotropy ◽  
Single Master Curve ◽  
Fluid Transfer

The anisotropic properties of clay-rich porous media have significant impact on the directional dependence of fluids migration in environmental and engineering sciences. This anisotropy, linked to the preferential orientation of flat anisometric clay minerals particles, is studied here on the basis of the simulation of three-dimensional packings of non-interacting disks, using a sequential deposition algorithm under a gravitational field. Simulations show that the obtained porosities fall onto a single master curve when plotted against the anisotropy value. This finding is consistent with results from sedimentation experiments using polytetrafluoroethylene (PTFE) disks and subsequent extraction of particle anisotropy through X-ray microtomography. Further geometrical analyses of computed porous media highlight that both particle orientation and particle aggregation are responsible of the evolution of porosity as a function of anisotropy. Moreover, morphological analysis of the porous media using chord length measurements shows that the anisotropy of the pore and solid networks can be correlated with particle orientation. These results indicate that computed porous media, mimicking the organization of clay minerals, can be used to shed light on the anisotropic properties of fluid transfer in clay-based materials.

Spherical Wave Propagations of Multiple Energy Sources in Fluid-Saturated Elastic Porous Media

2011 ◽  
Author(s):  
Lu Han ◽  
Liming Dai
Keyword(s):  
Porous Media ◽  
Spherical Wave ◽  
Three Dimensional ◽  
Energy Resources ◽  
Energy Sources ◽  
Multiple Point ◽  
Elastic Space ◽  
Compressive Wave ◽  
Point Energy ◽  
Wave Superposition

This research focuses on quantitative analysis of elastic space waves propagating porous media saturated with fluid. Compressive wave propagations are described in three dimensional (3-D) spherical coordinates in terms of displacements of the fluid and solid of the porous media. The wave superposition properties under multiple energy resources are investigated. Relative displacements between the fluid and solid of a porous medium in 3-D domain are quantified with consideration of the wave propagations excited by multiple point energy sources. Numerical analyses are performed and practically sound results are obtained.

scholarly journals The effect of training image and secondary data integration with multiple-point geostatistics in groundwater modelling

2014 ◽  
Vol 18 (8) ◽  
pp. 2943-2954 ◽  
Author(s):  
X. L. He ◽  
T. O. Sonnenborg ◽  
F. Jørgensen ◽  
K. H. Jensen
Keyword(s):  
High Resolution ◽  
Three Dimensional ◽  
Secondary Data ◽  
Training Image ◽  
Geophysical Data ◽  
Geostatistical Simulation ◽  
Multiple Point ◽  
Multivariate Distributions ◽  
Stochastic Hydrogeology ◽  
Capture Zones

Abstract. Multiple-point geostatistical simulation (MPS) has recently become popular in stochastic hydrogeology, primarily because of its capability to derive multivariate distributions from a training image (TI). However, its application in three-dimensional (3-D) simulations has been constrained by the difficulty of constructing a 3-D TI. The object-based unconditional simulation program TiGenerator may be a useful tool in this regard; yet the applicability of such parametric training images has not been documented in detail. Another issue in MPS is the integration of multiple geophysical data. The proper way to retrieve and incorporate information from high-resolution geophysical data is still under discussion. In this study, MPS simulation was applied to different scenarios regarding the TI and soft conditioning. By comparing their output from simulations of groundwater flow and probabilistic capture zone, TI from both sources (directly converted from high-resolution geophysical data and generated by TiGenerator) yields comparable results, even for the probabilistic capture zones, which are highly sensitive to the geological architecture. This study also suggests that soft conditioning in MPS is a convenient and efficient way of integrating secondary data such as 3-D airborne electromagnetic data (SkyTEM), but over-conditioning has to be avoided.

scholarly journals Challenges in Three-Dimensional Printing of High-Conductivity Copper

2018 ◽  
Vol 140 (2) ◽  
Author(s):  
Tahany I. El-Wardany ◽  
Ying She ◽  
Vijay N. Jagdale ◽  
Jacquelynn K. Garofano ◽  
Joe J. Liou ◽  
...  
Keyword(s):  
Induction Motor ◽  
High Efficiency ◽  
Three Dimensional ◽  
Surface Oxidation ◽  
Electric Machine ◽  
Optimum Process ◽  
Thin Wall ◽  
Stator Core ◽  
Direct Energy ◽  
Complex Features

With recent advancements in additive manufacturing (AM) technology, it is possible to deposit copper conductive paths and insulation layers of an electric machine in a selective controlled manner. AM of copper enables higher fill factors that improves the internal thermal conduction in the stator core of the electric machine (induction motor), which will enhance its efficiency and power density. This will reduce the motor size and weight and make it more suitable for aerospace and electric vehicle applications, while reducing/eliminating the rare-earth dependency. The objective of this paper is to present the challenges associated with AM of copper coils having 1 × 1 mm cross section and complex features that are used in producing ultra-high efficiency induction motor for traction applications. The paper also proposes different approaches that were used by the authors in attempts to overcome those challenges. The results of the developed technologies illustrate the important of copper powder treatment to help in flowing the powder easier during deposition. In addition, the treated powder has higher resistance to surface oxidation, which led to a high reduction in porosity formation and improved the quality of the copper deposits. The laser powder direct energy deposition (LPDED) process modeling approach helps in optimizing the powder deposition path, the laser power, and feed rate that allow the production of porosity free thin wall and thin floor components. The laser powder bed fusion (LPBF) models identify the optimum process parameters that are used to produce test specimens with >90% density and minimum porosity.

scholarly journals The effect of training image and secondary data integration with multiple-point geostatistics in groundwater modeling

2013 ◽  
Vol 10 (9) ◽  
pp. 11829-11860 ◽  
Author(s):  
X. He ◽  
T. O. Sonnenborg ◽  
F. Jørgensen ◽  
K. H. Jensen
Keyword(s):  
High Resolution ◽  
Groundwater Modeling ◽  
Three Dimensional ◽  
Secondary Data ◽  
Training Image ◽  
Geophysical Data ◽  
Multiple Point ◽  
Multivariate Distributions ◽  
Stochastic Hydrogeology ◽  
Airborne Electromagnetic Data

Abstract. Multiple-point geostatistic simulation (MPS) has recently become popular in stochastic hydrogeology, primarily because of its capability to derive multivariate distributions from the training image (TI). However, its application in three dimensional simulations has been constrained by the difficulty of constructing 3-D TI. The object-based TiGenerator may be a useful tool in this regard; yet the sensitivity of model predictions to the training image has not been documented. Another issue in MPS is the integration of multiple geophysical data. The best way to retrieve and incorporate information from high resolution geophysical data is still under discussion. This work shows that TI from TiGenerator delivers acceptable results when used for groundwater modeling, although the TI directly converted from high resolution geophysical data leads to better simulation. The model results also indicate that soft conditioning in MPS is a convenient and efficient way of integrating secondary data such as 3-D airborne electromagnetic data, but over conditioning has to be avoided.

scholarly journals Locality-based 3-D multiple-point statistics reconstruction using 2-D geological cross sections

2018 ◽  
Vol 22 (12) ◽  
pp. 6547-6566 ◽  
Author(s):  
Qiyu Chen ◽  
Gregoire Mariethoz ◽  
Gang Liu ◽  
Alessandro Comunian ◽  
Xiaogang Ma
Keyword(s):  
Cross Sections ◽  
Three Dimensional ◽  
Training Image ◽  
Sensitivity Analyses ◽  
Multiple Point ◽  
Multiple Point Statistics ◽  
Training Images ◽  
Critical Issues ◽  
Probability Aggregation ◽  
Novel Strategy

Abstract. Multiple-point statistics (MPS) has shown promise in representing complicated subsurface structures. For a practical three-dimensional (3-D) application, however, one of the critical issues is the difficulty in obtaining a credible 3-D training image. However, bidimensional (2-D) training images are often available because established workflows exist to derive 2-D sections from scattered boreholes and/or other samples. In this work, we propose a locality-based MPS approach to reconstruct 3-D geological models on the basis of such 2-D cross sections (3DRCS), making 3-D training images unnecessary. Only several local training subsections closer to the central uninformed node are used in the MPS simulation. The main advantages of this partitioned search strategy are the high computational efficiency and a relaxation of the stationarity assumption. We embed this strategy into a standard MPS framework. Two probability aggregation formulas and their combinations are used to assemble the probability density functions (PDFs) from different subsections. Moreover, a novel strategy is adopted to capture more stable PDFs, where the distances between patterns and flexible neighborhoods are integrated on multiple grids. A series of sensitivity analyses demonstrate the stability of the proposed approach. Several hydrogeological 3-D application examples illustrate the applicability of the 3DRCS approach in reproducing complex geological features. The results, in comparison with previous MPS methods, show better performance in portraying anisotropy characteristics and in CPU cost.

Sign in / Sign up

Export Citation Format

Share Document