Permeability of Uniformly Graded 3D Printed Granular Media

Abstract

The present work explores water permeability of uniformly graded irregular grains using 3D printing with controlled shapes and fractal morphological features at low Reynold's number for viscous flow. From large amount of real 3D granular morphological data, a scaling law, in terms of fractal dimension, is found to be followed. With this universal law, sand grains with controlled fractal morphological features are generated using Spherical Harmonics, and then created using 3D printing technique for water permeability tests. A modified Kozeny‐Carman equation is proposed through more accurate determination of specific area, as a function of relative roughness and fractal dimension, than approximation using the volume‐equivalent sphere. By isolating the contributions from specific area, the shape coefficient is found to be insensitive to particle morphology. Through benchmarking the model prediction against experiments from both this work and past literature, we demonstrate the validity and wide applicability of the modified Kozeny‐Carman equation.

Investigation of Flow Characteristics of Landslide Materials Through Pore Space Topology and Complex Network Analysis

“…Samples 7∼10 have the same GSD and porosity but different grain shapes. We can see that the permeability decreases with increasing shape irregularity, which is consistent with previous findings (Garcia et al., 2009; Wei et al., 2021). Compared with grain size distribution, the influence of grain shape on flow permeability is relatively small and can be ignored in this study to highlight the effect of GSD.…”

“…The high permeability area of the loose dam body will increase the infiltration, which could cause seepage failure immediately (Shi et al, 2015).Previous studies on the hydraulic properties of landslide materials have revealed that the grain size distribution of the dam accumulation has significant impact on the seepage stability of landslide dams (Okeke & Wang, 2016;Zhu et al, 2020). The effect of grain shapes on permeability has also been investigated via laboratory tests (Wei et al, 2021) and numerical simulations (Garcia et al, 2009;Torskaya et al, 2014). The relationship between pore structure and hydraulic properties of landslide materials has not been fully understood, mainly due to the challenges in characterizing the pore structure of landslide materials.…”

Predicting 3D Physical Properties From a Single 2D Slice Based on Convolutional Neural Networks: 2D‐Slice‐To‐3D‐Properties for Porous Rocks

“…The permeability of hyperpermeable samples can be satisfactorily estimated by theoretical models (such as the modified Kozeny‐Carman equation or other reported models) using pore size or grain size data, thus not requiring the construction of complex networks (J.‐P. Wang et al., 2017; D. Wei et al., 2021). Thus, abnormal data with permeability >2 pixel 2 and formation factor >1,000 were excluded from the original data set.…”

“…For permeability, since this variable usually spans more than four orders of magnitude, in the normalization operation of data processing, in order to avoid the influence of high-permeability values on low-permeability samples, we artificially deleted samples with ultra-high permeability (larger than 2 pixel 2 ). The permeability of hyperpermeable samples can be satisfactorily estimated by theoretical models (such as the modified Kozeny-Carman equation or other reported models) using pore size or grain size data, thus not requiring the construction of complex networks (J.-P. Wang et al, 2017;D. Wei et al, 2021).…”

Time‐Fractional Flow Equations (t‐FFEs) to Upscale Transient Groundwater Flow Characterized by Temporally Non‐Darcian Flow Due to Medium Heterogeneity

“…Marsily, 1997), and the stochastic method (which assumes medium properties to be random variables) is generally based on assumptions not always valid for natural media (Dagan, 1993; Dykaar & Kitanidis, 1992; Gelhar & Axness, 1983). Recent studies also developed experimental techniques (Cai et al., 2012; Kendrick et al., 2021; Strangfeld, 2020; Wei et al., 2021), hydraulic tomography (Bellin et al., 2020), deep learning (Moghaddam, 2020), as well as multiscale characterization (Colecchio et al., 2020), among other approaches, to define the effective hydraulic conductivity. Here we select the Monte Carlo simulation of groundwater flow to calculate the effective hydraulic conductivity K e for two reasons.…”