0% found this document useful (0 votes)

18 views27 pages

4-Evaluation of Dynamic Analysis

Uploaded by

pranav1312002

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

18 views27 pages

4-Evaluation of Dynamic Analysis

Uploaded by

pranav1312002

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 27

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/337607681

Evaluation of Dynamic Image Analysis for Characterizing Granular Soils

Article in Geotechnical Testing Journal · November 2019

DOI: 10.1520/GTJ20190137

CITATIONS READS

27 4,281

2 authors:

Linzhu Li Magued Iskander

New York University New York University, Tandon School of Engineering
13 PUBLICATIONS 121 CITATIONS 243 PUBLICATIONS 4,197 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Linzhu Li on 28 November 2019.

The user has requested enhancement of the downloaded file.

Geotechnical
Testing Journal
Linzhu Li1 and Magued Iskander2

DOI: 10.1520/GTJ20190137

Evaluation of Dynamic Image

Analysis for Characterizing
Granular Soils

Copyright by ASTM Int'l (all rights reserved); Thu Nov 28 13:28:24 EST 2019
Downloaded/printed by
New York University (New York University) pursuant to License Agreement. No further reproductions authorized.
Geotechnical Testing Journal

doi:10.1520/GTJ20190137 available online at www.astm.org

Linzhu Li1 and Magued Iskander2

Evaluation of Dynamic Image Analysis

for Characterizing Granular Soils

Reference
L. Li and M. Iskander, “Evaluation of Dynamic Image Analysis for Characterizing Granular Soils,”
Geotechnical Testing Journal https://doi.org/10.1520/GTJ20190137

ABSTRACT
Manuscript received April 20, This study investigates the efficacy of dynamic image analysis (DIA) for determining particle size
2019; accepted for publication and shape distribution. The method employs a high-frame-rate camera to image individual
August 21, 2019; published online
November 25, 2019.
particles of sand that have been transported and separated using a stream of pressurized
air. DIA can generate both particle size and shape information and provides a quantitative stat-
1
Civil & Urban Engineering
istical description of the grain size and shape distribution within the specimen. The feasibility,
Department, NYU Tandon School
of Engineering, Six Metrotech repeatability, and accuracy of DIA for routine analysis of particle size and shape distribution was
Center, Brooklyn, NY 11201, USA investigated using 16 granular soils spanning a number of common sizes and shapes. Several
2 particle shape descriptors were evaluated, including aspect ratio, convexity, and sphericity.
Civil & Urban Engineering
Department, NYU Tandon School The effect of a variety of test parameters including moisture content, sample weight, primary
of Engineering, New York air pressure, and test duration were explored to determine the optimal specimen weight and
University, Six Metrotech Center,
equipment settings for DIA. Finally, the efficacy of DIA in resolving mixtures of fine and coarse
Brooklyn, NY 11201, USA
(Corresponding author), e-mail: sands was also explored. The method proved to be feasible, repeatable, and accurate for provid-
[email protected], https:// ing particle size distributions spanning four orders of magnitude, in terms of particle size. DIA
orcid.org/0000-0001-8245-1451 offers a number of advantages; the method is quick, requires small specimen sizes, and provides
quantitative information on approximately 3–4 % of the particles in the specimen.

Keywords
equivalent projected area of a circle, Feret diameter, number distribution, volume distribution,
round, silica, angular, quartz, sand, gap graded

Nomenclature
A = particle area
AR = aspect ratio, dFmin/dFmax
Cc = coefficient of gradation, D230/D60*D10
Cu = uniformity coefficient, D60/D30

Copyright by ASTM
Copyright © 2019Int'l (all rights
by ASTM reserved); Thu
International, 100 Nov
Barr 28 13:28:24
Harbor EST
Drive, PO2019
Box C700, West Conshohocken, PA 19428-2959
Downloaded/printed by
New York University (New York University) pursuant to License Agreement. No further reproductions authorized.
LI AND ISKANDER ON DYNAMIC IMAGE ANALYSIS OF GRANULAR SOILS

Cx = convexity = Ap/(Ap + As)

de = diameter of an equivalent circle having an area equal to that of the projected
particle area
dFmax = maximum Feret diameter, or, Feret-max
dFmin = minimum Feret diameter, or, Feret-min
Dxx = diameter corresponding to xx% of the specimen (by weight) that is smaller than xx
in microns (or, % passing)
P = particle perimeter
pffiffiffiffiffiffi
S = sphericity = 2 πA=P

Introduction
In recent years, the effect of particle size and shape distribution on the strength, void ratio, friction angle, and hydraulic
conductivity of granular soils has been increasingly recognized (e.g., Cho, Dodds, and Santamarina 2006; Rousé,
Fannin, Shuttle 2008; Bareither et al. 2008; Shin and Santamarina 2013; Zheng and Hryciw 2016a). Several techniques
have been suggested to measure the physical characteristics of soil particles. Pentland (1927) was perhaps the first to
suggest that granular soil behavior can be related to the constituent grain shape. Wadell (1932), Riley (1941), and
Krumbein (1941) were among the pioneers who developed approaches to define the form and roundness of particles.
A number of terms have been introduced to define particle shapes, including equivalent circle diameter, Feret value,
aspect ratio (AR), sphericity (S), convexity (Cx), and roundness (Table 1).
One of the most commonly used techniques for obtaining particle shape information is microscopy.
Microscopy provides information for determining particle size and shape for particles as small as 0.003 μm.
However, the technique remains cumbersome and difficult to apply for conveniently determining particle size
and shape distribution. Thus, obtaining statistically significant data using microscopy can be time consuming
because image segmentation for particles is difficult for densely crowded particles, making the process difficult to
automate at this time. Presently, automatic image identification and segmentation are only possible for images
with well-spaced individual particles. Microscopy images, on the other hand, typically consist of particles over-
lapping each other, so the determination of shape descriptors is challenging to automate.

TABLE 1
Particle shape descriptors.

Descriptor Formula Description Reference

EQPC de Equivalent circle diameter (ASTM F1877-16)

Feret’s value (F) dFeret Feret diameter (Kuo and Freeman 2000)
Aspect ratio (AR)a dFmin/dFmax The ratio between the minimum feret’s (ISO 9276-6 2008)
diameter and the maximum feret’s diameter
Convexity (Cx) Ap/(Ap+As) The ratio between the projection particle (Mora and Kwan 2000)
area and the area of the convex hull

Sphericity (S)a pffiffiffiffiffiffi Smoothness of the perimeter (Altuhafi, O’Sullivan, and
2 πA P
Cavarretta 2013)
Roundness X
N A description of the presence of irregularities (Wadell 1932)
ri N
R= i=1
rins

Pictorial representation of notation

Note: a AR is also defined as S according to Krumbein and Sloss chart (1951), Zheng and Hryciw (2016b), and others. However, in this study, AR and S
are defined consistently according to this figure, and its cited references.

Copyright by ASTM Int'l (all rights reserved); Thu Nov 28 13:28:24 EST 2019 Testing Journal
Geotechnical
Downloaded/printed by
New York University (New York University) pursuant to License Agreement. No further reproductions authorized.
LI AND ISKANDER ON DYNAMIC IMAGE ANALYSIS OF GRANULAR SOILS

Particle size distribution is the most commonly employed property of granular soils. Many standards and codes
have incorporated particle size as a parameter in soil classification. Conventional standard sieve analysis such as
ASTM D6913/6913M (2017), Standard Test Methods for Particle-Size Distribution (Gradation) of Soils Using Sieve
Analysis, was the first method developed for determination of size distribution and still remains popular today
with many standards in use (e.g., BS 1377-1 (2016), Methods of Test for Soils for Civil Engineering Purposes.
General Requirements and Sample Preparation; ISO 11277 (2009), Soil Quality – Determination of Particle Size
Distribution in Mineral Soil Material – Method by Sieving and Sedimentation; JIS A 1204 (2009), Test Method
for Particle Size Distribution of Soils; JTG E40 (2007), Test Methods of Soils for Highway Engineering; DIN ISO
11277:2002, Soil Quality - Determination of Particle Size Distribution in Mineral Soil Material - Method by
Sieving and Sedimentation). These standards are relatively similar (Table 2). However, until recently, use of
the sieve analysis to quantify particle size limited the accuracy and scope of particle size classification. As a result,
visual classification and engineering judgment played an important role in the analysis, at least on a practical level.
Consequently, the results tend to be subjective, case specific, and varied among different observers. Although
standard sieve analysis is reliable and popular, it has a number of limitations. Chiefly, the test does not provide
any information on particle shape. Additionally, precision is limited, especially for smaller particle sizes. Finally, the
equipment is noisy, dusty, and generates vibrations that may affect operations of nearby sensitive laboratory
equipment.
Laser diffraction was recently introduced for measuring particle size distribution in order to eliminate
some of the drawbacks of conventional sieve analysis, especially for fine and very fine particles. Laser dif-
fraction is based on the principle that different size particles diffract light at different angles, such that coarser
particles scatter light at smaller angles than finer ones. Thus, the particle size is illustrated by the light scatter-
ing pattern (Polakowski et al. 2014). In this technique, the particle size distribution is reported as the dis-
tribution of spherical particles with the same diffraction pattern as the measured particles. In that regard, the
results represent equivalent sphere diameter (Yu and Hancock 2008). Laser diffraction permits obtaining the
particles size distribution of a large number of particles conveniently, especially for very fine particles (as small
as 0.1 μm). However, because it assumes that all particles are spherical in shape, it may not always be rea-
sonable for angular materials.
In recent years, dynamic image analysis (DIA) has been introduced to measure particle size and shape
distribution, especially in the pharmaceutical industry (e.g., Yu and Hancock 2008; Nalluri et al. 2010). DIA
can generate both particle size distribution and particle shape information more conveniently than any other
technique. In addition, DIA can provide many more descriptors of particle shape than laser diffraction or sieve
analysis, and comparable to microscopy. DIA typically employs a high air pressure (up to 5 bar [500 kPa]) to

TABLE 2
Comparison of standard sieve sizes standards in typical ranges

American ASTM E11-17 International aISO 3310 bISO 565 China T 0115-1993 Japan JIS Z 8801-1:2006

Sieve No. Opening (mm) Opening (mm) Sieve No. Opening (mm) Sieve No. Opening (mm)

4 4.75 4.75 4 4.75 4 4.75

10 2.00 2.00 10 2.00 8.6 2.00
20 0.85 0.85 18 1.00 18 0.850
40 0.425 0.425 35 0.500 36 0.425
60 0.250 0.250 60 0.250 60 0.250
100 0.150 0.150 … … … …
140 0.106 0.106 … … 140 0.106
200 0.075 0.075 200 0.075 200 0.075

Note: a BS 401-1: 2000 was replaced by BS ISO 3310-1: 2016, Test Sieves – Technical Requirements and Testing – Part 1: Test Sieves of Metal Wire Cloth;
DIN 4188-1: 1977-10 was replaced by DIN ISO 3310-1: 2001-09; b NFX11 501: 1970 was superseded by NF ISO 565 1990, this standard was last
reviewed and confirmed in 2012.

transport and separate particles, and images of sparse particles are taken with a high-frame-rate camera. DIA reliably
overcomes the issue of the overlapping particle in microscopy, while permitting processing a large volume of soil
similar to sieve analysis. DIA also provides a quantitative statistical description of the grain size distribution.
DIA has been recently adopted in geotechnical engineering and was used for creating a research database of 36
sands with quantified size and shape information that was used for exploring the essential relationship between par-
ticle shape and mechanical properties (Altuhafi and Coop 2011; Altuhafi, O’Sullivan, and Cavarretta 2013; Altuhafi,
Coop, and Georgiannou 2016; Yang and Luo 2015). However, detailed requirements for conducting DIA for routine
geotechnical characterization are not available. The aim of this study is to investigate the feasibility, repeatability, and
accuracy of DIA for routine analysis of particle size and shape distribution. Sixteen granular soils were selected for
testing, including two different particle shapes: round silica and angular quartz sand. Several particle shape descriptors
were evaluated in comparison to conventional descriptors, such as AR, Cx, and S. Conventional grain size distribution
sieve analysis was also compared with different particle shape descriptors, including Feret value and equivalent pro-
jected area of a circle (EQPC). The effects of different test parameters, such as moisture content, sample weight,
primary air pressure, and test duration, were investigated to determine optimal specimen and equipment specifications
for DIA. Finally, the efficacy of DIA in resolving mixtures of fine and coarse sands was also explored.

Materials Tested
Sand particles with two different shapes, round and angular, were selected for investigation. Sixteen commercially
sourced sands were employed throughout this study and identified by their commercial names. The selected
particle grain size mainly ranges from 75 μm to 2000 μm. These sizes were selected to be in compliance with
ASTM’s Unified Soil Classification System definition of sand (Howard 1984). The sand used in this experiment is
further divided into two categories, rounded and angular, as follows:

• Rounded silica sands, denoted as #12-20, #20-30, #40-70, #50-80, #70-100, #100-140, #140-200, #140-270.
The denotation refers to the passing and retaining sieve sizes.
• Angular quartz sands, denoted as #4, #3, #2, #1, # 12 , # 20 , # 30 , # 40 , in decreasing size order.

Scanning electron microscopic (SEM) images of all sands are shown in figure 1. Both sands have a specific
gravity of 2.65 and hardness of Mohs-7. More than 99.8 % of silica sand and 99.5 % of quartz sand is made of
silicon dioxide (SiO2). The main difference, other than particle shape, is that quartz contains mineral impurities
such as titanium, iron, manganese, making it appear more transparent compared with silica sand. The SEM
images demonstrate that the grain shapes of quartz sand are sharper than those of silica sand and that there
is a decrease of silica sand particle roundness with the change of particle size.

Background
OVERVIEW OF THE DIA APPARATUS
A QicPic device was employed in this study. The device consists of a vibratory feeding system known as VIBRI,
a dispersing system known as RODOS, and the QicPic DIA sensor (fig. 2). The device operates as follows:
(a) samples are poured into the VIBRI device. The device consists of a hopper capable of feeding sand at a con-
stant rate of mass flow in order to provide precise dosing of the tested material into the RODOS dispersing unit.
(b) The RODOS disperses the fed material using compressed air, creating a well-dispersed particle stream through
an image plane. RODOS employs a 4-mm-wide straight dispersing line, resulting in particles traveling within a
finite depth in order to eliminate the blur caused by of out-of-focus images. The dispersed flow causes particles to
travel separately from each other, in order to minimize overlapping of particles. (c) The QicPic images the indi-
vidual particles as they travel. Particle shapes are captured at a frame rate of up to 500 frames per second using
a high-speed camera having a 4-megapixel (2,336 by 1,728) resolution. Different camera lenses are available to
cover different particle size measuring ranges. The selection of the proper lens is critical for obtaining high-quality

FIG. 1 SEM images of quartz and silica sands with their magnification factor.

Quartz #1 20× Quartz #2 20× Quartz #3 20× Quartz #4 20×

Silica #40-70 20× Silica #35-50 20× Silica #20-30 20× Silica #12-20 20×

Quartz #4/0 100× Quartz #3/0 40× Quartz #2/0 20× Quartz #1/2 20×

Silica #140-270 20× Silica #100-140 20× Silica #70-100 20× Silica #50-80 20×

FIG. 2
QicPic apparatus
including feeding system
(VIBRI), dispersing
system (RODOS), and
DIA sensor (QICPIC).

images with high contrast in order to capture sharp particle edges. In this study, the lens M7 with the measuring
range of 4 μm to 2,888 μm was employed. A special pulsed laser combined with an adaptable beam expander is
used to permit an extremely low exposure time to prevent motion blur of moving objects. The pulsed light source
provides illumination time of less than 1 nanosecond to prevent motion blur, because 100 nm of motion blur
corresponds to less than 1 μm, which is smaller than the size of one pixel (Nalluri et al. 2010). Edges are precisely
detected when the particle flow is imaged in transmission using a special objective lens transmitting light rays to
the camera that are parallel to the optical axis. Even though some particles, such as quartz, are highly transparent,
they appear as black in the captured images. The device employs a powerful imaging algorithm, which can evalu-
ate millions of particles in minutes and ensure statistically relevant results. The particle analysis outcome includes
values for grain size distribution and grain shape descriptors.

GRANULAR SOIL PARTICLE SIZE AND SHAPE DESCRIPTORS

All grain size measurements typically attempt to convert three-dimensional (3-D) bodies into size (1-D) or shape
(2-D) descriptors. This process is influenced by the particle orientation during measurement, which is test de-
pendent. DIA can capture and calculate the projected (i.e., 2-D) area of many, but not all, particles in the video
stream. The mean shape of 2-D images is believed to correlate reasonably well with 3-D shapes, if a sufficient
number of particles with ample pixel resolution is analyzed, which can be achieved using the current generation of
QicPic for particles larger than 40 μm (Cepuritis et al. 2017).
One of the most common shape descriptors in geotechnical engineering is Roundness, which is defined as
the ratio of the average radius of curvature of the corners of a particle to the radius of the maximum inscribed
circle as shown in Table 1 (Wadell 1932, 1933). Roundness is not directly provided by DIA. Because the original
procedures suggested by Wadell for determining the radii of curvature in a particle and normalizing them by the
radius of the largest inscribed circle (Table 1), roundness is typically evaluated using charts developed by
Krumbein and Sloss (1951) or Powers (1953). These charts provide a quantitative and qualitative estimation
of particle angularity and S ranging from angular to well-rounded and low S to high S, respectively.
Although these charts help classify soil particle shapes, they are subjective and require much manual effort.
The state of the art in determining roundness is to employ computational geometry techniques for measuring
roundness (Zheng and Hryciw 2015, 2016b).
Synchrotron microcomputed tomography has been used to acquire high-resolution images of particles
of sand and characterize their morphological properties in 3-D (Alshibli et al. 2015). However, at this time
3-D measurements are too cumbersome and expensive for routine geotechnical practice. DIA, on the contrary,
captures 2-D shape in a short period of time, while sieve analysis and laser diffraction only characterize 1-D size.
Its results include the number of imaged particles and 2-D projection of all the particles imaged in the video
stream. The results can be presented in terms of size descriptors or shape descriptors, as follows.

Size Descriptors
Size descriptors can be presented in terms of a number distribution, length distribution, particle area distribution,
or volume distribution, as follows:
Number distribution typically refers to the percentage of the number of particles that is equal or smaller than
a certain size. The value of D50 corresponding to the number distribution is always remarkably smaller than
the D50 obtained by conventional mass distribution (sieve analysis) because the number distribution is typically
dominated by the fine particles, whereas the mass distribution is usually dominated by the coarse fraction of the
tested specimen.
Length (or size) distribution may be presented as the diameter of a circle having an area equivalent to that
of the projected 2-D particle, known as equivalent projected area of a circle (EQPC or de). Although EQPC is
widely used for evaluating the shapes of many particles, it is not optimal for angular particles. EQPC typically
results in a coarser particle size distribution than what would be measured using conventional sieve analysis,
because grain size distributions obtained in sieve analysis are skewed. Several definitions of soil particle size

and shape descriptors are summarized in Table 1. Feret diameters are commonly employed for describing
measured dimensions of particles. The Feret diameter refers to the distance between two parallel tangents
to the particle at an arbitrary angle (Kuo and Freeman 2000). Thus, for any particle, an unlimited number of
Feret diameters can be computed as the particle is rotated between two parallel lines. For irregular particles, the
Feret-max (dFmax) and Feret-min (dFmin) are the largest and smallest values resulting from this operation,
corresponding to the longest and shortest distance between two tangents. Feret-max and Feret-min are directly
supplied by DIA. In practice, Feret-min is often used as the diameter equivalent to that obtained in sieve analy-
sis (e.g., White 2003).
Volume (or mass) distribution is computed by calculating the volume of each particle, assuming all particles
to be spheres. Sphere’s diameter is typically EQPC but can also be taken as dFmax or dFmin. The volume calculation
from DIA and laser diffraction are therefore similar, except that DIA permits use of multiple diameters as
a reference. If (1) a given specific gravity value of the tested soil sample is known, or assumed, and (2) all particles
have the same specific gravity; the corresponding particle size distribution from DIA is consistent with the mass
distribution obtained from sieve analysis, for dFmin.
To study the impact of different particle shape descriptors on the resulting particle size distribution, two
sands were selected for comparison, rounded silica #20-30 and angular quartz #3. The grain size distributions
obtained using sieve analysis (ASTM D6913/D6913M) as well as volume distribution obtained using various
DIA reference diameters (EQPC, dFmax, dFmin) are plotted in figure 3. As expected, dFmin is the closest distribution
to that obtained from sieve analysis. Feret diameters of Quartz #3 vary more than that of Silica #20-30 because of
the effect of its angularity on the Feret measurement.
For both sands, the distribution obtained using EQPC fell in between the two distributions obtained
using maximum and minimum Feret diameters. The dFmax diameter deviates the most from the traditional sieve
analysis and appears to be least useful, because sieve analysis has established itself as the reference grain size
distribution, at least in geotechnical engineering practice.
It is interesting to note that DIA reveals errors in particle size distributions obtained through standard
sieve analyses. For example, the particle size distribution of quartz sand shown at figure 3 is markedly
different than that obtained through DIA, for particle sizes smaller than 250 μm (Sieve No. 60). In particular,
when EQPC or Feret-min shape descriptors are employed, the particle size distribution is close to that ob-
tained in sieve analyses for sizes larger than 250 μm, but DIA and sieve diverge below that size, because sieve
analysis lacks the resolution to capture the correct grain size distribution of fine soils. For example, 0 g of
Quartz #3 was collected on bottom pan and sieve #200, but small amounts of material could clearly be iden-
tified using DIA.

FIG. 3
Impact of size
descriptors EQPC and
Feret Diameter on the
particle size distributions
for (A) Silica #20-30 and
(B) Quartz #3.

Shape Descriptors
DIA provides information on particle morphology including AR, S, Cx, elongation, and straightness. AR, S,
and Cx are most commonly used for sedimentary materials. These parameters are illustrated graphically for
Silica #20-30 and Quartz #3 in figure 4, while elongation and straightness are employed to characterize fibers.
These 2-D shape morphology descriptors are described next:
AR is defined as the ratio of the minimum to the maximum Feret diameters (Table 1), AR = dFmin/dFmax
(ISO 9276-6 2008). The AR ranges from 0 to 1, corresponding to a range of shapes, from extremely elongated
particles to perfectly rounded particles. AR captures the rounded nature of Silica #20-30 and the angularity of
Quartz #3 (fig. 4).
S describes the resemblance of an irregularly shaped soil particle to a perfect sphere, and there are
several definitions in current use. Altuhafi, O’Sullivan, and Cavarretta (2013) defined S as the ratio between
the perimeter of a circle, calculated using the EQPC diameter, de, and the perimeter, P, of the corresponding
pffiffiffiffiffiffi
projected area, as follows: S = 2 πA=P, where A = πd2e =4. Thus, S is proportional to the square root of particle
area and inversely proportional to particle perimeter, which is the definition employed in this study.
Other definitions of S have been summarized by Rodriguez, Edeskär, and Knutsson (2013), including:
(1) S = dFmin/dFmax (Wadell 1935; Krumbein and Sloss 1951) and (2) S = P/A, i.e., particle perimeter to area
(Janoo 1998).
Convexity (Cx) is used to describe the particles’ compactness. It is the ratio between the projection particle
area (Ap) and the area of the convex hull (Ap+As), as shown in Table 1 (Mora and Kwan 2000). The convex hull is
the area of the minimum convex boundary circumscribing the particle. Convexity represents how close particle
can be expressed as the convex hull, a measure that is intended to capture the surface morphology, i.e., degree of
roughness or presence of asperities, in a particle.
The distributions of AR, Cx, and S obtained using DIA for both Silica #20-30 and Quartz #3 are pre-
sented in figures 4–7, where a 10-g specimen of each sand was employed. This amount corresponded to an
average of 523 and 957 individual particles of Silica #20-30 and Quartz #3, respectively. AR, Cx, and S, are
illustrated graphically for the two test sands in figure 4. The distribution of the morphology parameters
within each sand is evident in Table 3, where the number of particles corresponding to nine shape intervals,
are tabulated for AR, Cx, and S. The mean size D50 corresponding to each interval is also computed based on
the number and volume distribution within each sand. The corresponding image of a particle representing

FIG. 4
Representative particle
shapes corresponding to
morphology parameters
(1) AR, (2) Cx, and (3) S
of Quartz #3 and Silica
#20-30 sands.

FIG. 5 Comparison of particle shape descriptors by number and volume for Silica #20-30 and Quartz #3. (A) S, (B) AR,
and (C) Cx. All specimens were 10 g, air-dried, and tested at a primary air pressure of 3 bar (300 kPa).

D50 is selected from the video stream for inclusion in figure 4 (based on volume distribution of EQPC). It is
noteworthy that for Quartz #3 D50 corresponding to the volume distribution is significantly larger than
that corresponding to the number distribution owing to the fact that the number distribution is dominated
by fine particles, i.e., there are significantly more smaller particles than larger particles. Silica #20-30 is
poorly graded, with little fines, thus, volume versus number distribution was hardly influenced by presence
of fines.
The density distribution of S, AR, and Cx is plotted in figure 5, for both sands, in terms of both frequency
(histogram) distributions of that particle shape in terms of both number and volume (i.e., mass) distributions.
The markers represent the center of bin sizes used for analysis. All distributions are somewhat skewed. The
number and volume distributions of Silica #20-30 were similar for all shape parameters, owing to the uni-
formity of the product. There is less than 1 % difference between the number and volume distributions at all
intervals, owing to Silica #20-30 being a commercial product with a relatively uniform size and shape dis-
tribution within each size interval. For Quartz #3, presence of fines had a marked effect on the distribution
of Cx and S, but not on AR. The comparison of the effect of number and volume distribution on AR fluctuates
across the AR ranges, depending on whether large or small particles dominate or are equal in any interval of the
AR range. A higher density was observed for Quartz #3 at S = 0.7–0.8 using the volume distribution in

FIG. 6 Correlation of particle shape descriptors for Silica #20-30 and Quartz #3 sands (A) S and Cx; (B) S and AR; (C) AR
and Cx. All specimens were 10 g, air-dried, and tested at a primary air pressure of 3 bar (300 kPa).

comparison to the number distribution. The total number of particles in this sphericity interval is 42 %, but the
particle volume accounts for 63 % of the total volume. This suggests that most large particles possess this shape
interval.
S is plotted against AR or Cx for each particle in figure 6. In general, silica exhibited a narrow shape
parameter interval and higher maximum values, thus, they plot at the top right corner in all figures. As ex-
pected, quartz, because of its angularity, exhibited a wider range in the measured values than silica. For ex-
ample, for quartz #3, S = 0.24–0.88, AR = 0.15–0.94, and Cx = 0.47–0.98 with the skewness of −0.82, −0.31,
and −1.97 while for Silica #20-30, S = 0.47–0.93, AR = 0.42–0.97, and Cx = 0.84–0.99 with the skewness of
−0.79, −0.56, and −2.15, respectively. For both sands, the skewness numbers suggest that values of S, AR, and
Cx are skewed toward larger numbers. The kurtosis for S, AR and Cx of quartz is 0.61, −0.57, and 7.97, re-
spectively, and 0.23, 0.11, and 7.25 for silica, respectively. These kurtosis values suggest that S and AR follow a
normal distribution, whereas Cx distributions does not. Additionally, silica has a better Gaussian distribution
compared with quartz in terms of S and AR. Sphericity and convexity are somewhat correlated, with R2 = 0.68
and 0.57 for Silica #20-30 and Quartz #3, respectively. AR and S are not correlated (R2 = 0.02 for silica and 0.16
for quartz) as was AR and Cx (R2 = 0.01 for silica and 0.02 for quartz). For both sands, these results indicate

FIG. 7 Effect of particle size on (A) Cx, (B) S, and (C) AR for Silica #20-30 and Quartz #3 sands. All specimens were 10 g,
air-dried, and tested at a primary air pressure of 3 bar (300 kPa).

that AR and S are independent, thus, each can vary without affecting the other. This is not surprising and has
been pointed out previously by Krumbein and Sloss (1951) as well as Zheng and Hryciw (2016b). This study
demonstrates that AR and Cx are also independent, whereas S and Cx are somewhat related.
The Cx, S, and AR versus probabilistic particle size distribution (by volume) is illustrated in figure 7. The mean
values of the three shape descriptors (AR, S, Cx) of Silica #20-30 are consistently larger than those of Quartz #3 for
the majority of particle sizes in the particle size density distribution curve. In summary, the grain size distribution for
the largest 90 % of particles (by weight) is dominated by particles having S = 0.7–0.8 and Cx = 0.7–0.9 for both sands
whereas AR = 0.7–0.9 and 0.6–0.7 for Silica #20-30 and Quartz #3, respectively. These values are consistent with the
inferences obtained from figure 5. It is also interesting to note that while both Silica #20-30 and Quartz #3 exhibit a
Gaussian distribution with respect to size, none of the three shape descriptors appears to have a marked size depend-
ence. This may be understood in the context of fractal particle evolution, i.e., when particles fracture, daughter
particles may preserve a similar S, AR, and Cx to that of the parent material. This is not surprising because particle
size distributions of crushed granular materials have been found to be self-similar or fractal to the parent material,
under certain conditions (Turcotte 1986), and models to relate the particle scale to the fractal dimensions have been
proposed (McDowell, Bolton, and Robertson 1996). However, Quartz #3 exhibits a somewhat more stable S, AR, and
Cx, in comparison with Silica #20-30, especially for the finest fraction of the specimen. In fact, 80 % of particles have
S > 0.6 for Quartz #3 and S > 0.7 for Silica #20-30.

TABLE 3
Comparison of particle size D50 on number and volume distributions of different particle shape descriptors for Quartz #3 and
Silica #20-30

Particle Shape Evaluated

Sand Descriptor Diameter, μm 0.1–0.2 0.2–0.3 0.3–0.4 0.4–0.5 0.5–0.6 0.6–0.7 0.7–0.8 0.8–0.9 0.9–1.0

Quartz #3 AR Particle number 2 19 62 144 188 224 208 102 8

D50 number 610 291 277 234 222 220 208 211 234
D50 volume 610 1046 957 1,181 1,038 1,117 1,108 1,090 915
Convexity Particle number … … … 1 2 4 50 400 500
D50 number … … … 153 144 158 158 186 326
D50 volume … … … 153 72 216 395 988 1,133
Sphericity Particle number … 1 7 32 122 252 405 136 …
D50 number … 153 148 193 213 221 245 201 …
D50 volume … 153 155 213 304 387 1,166 1,038 …

Silica #20-30 AR Particle number … … … 2 15 84 164 205 53

D50 number … … … 771 809 808 808 820 823
D50 volume … … … 771 839 824 820 828 871
Convexity Particle number … … … … … … … 3 520
D50 number … … … … … … … 729 815
D50 volume … … … … … … … 729 826
Sphericity Particle number … … … 3 17 72 171 234 26
D50 number … … … 822 818 817 812 818 801
D50 volume … … … 822 826 824 826 830 812

Note: 1. The number of particles captured by DIA includes out-of-focus particles and dust sticking to the sands; as a result, the particle number shown
in Table 3 represents particles with an EQPC diameter larger than D1; 2. D50 number represents the EQPC corresponding to the median particle
EQPC, i.e., 50 % of the particles by number are larger/smaller than the indicated EQPC in microns; 3. D50 volume represents the EQPC corresponding
to 50 % of the specimen by volume (also weight) that is smaller than indicated EQPC in microns.

Effect of Various Test Parameters on Experimental Results

It is important to understand the effect of test parameters and specimen size on the repeatability of the shape and
morphology parameters measured using DIA. To do so, a battery of tests was conducted using four different
sands, in which the specimen weight, moisture content, and primary air pressure were varied. Silica #20-30
and Quartz #3 (D50 = 804 μm and 1,155 μm, respectively) were used in all tests. For specimen size effect,
Silica #70-100 and Quartz # 12 (D50 = 287 μm and 194 μm, respectively), were also employed in an attempt
to account for the correlation between sample weight, particle size, and gradation.

EFFECT OF PRIMARY AIR PRESSURE

DIA measures particle projections to gather both particle size and shape information. Consequently, a dispersed
particle flow is desired during the experiments. Primary air pressure is used to obtain well-separated particles. It
should be strong enough to separate agglomerated particles from each other but not damage or crush any particles.
The tested materials were not prone to breakage, so the study focused only on the necessary pressure to achieve a well-
dispersed particle flow. Several tests of Silica #20-30 and Quartz #3 were conducted at different primary air pressures
ranging from 1 to 5 bar (100 to 500 kPa), and the particle size distribution curves are shown in figure 8. The con-
sistency of the particle size distribution of Silica #20-30 at all 5 primary air pressures indicates that the change of air
pressure does not affect the particle size distribution. However, for Quartz #3, the grain size distribution shows a
coarser distribution at 1 bar than at other pressures because of the failure of that pressure to separate fine particles. In
particular, the value of D10 measured using 1 bar is 831 μm, which is different from the average value of 708 μm of the
other 4 tests. This phenomenon was not observed for Silica #20-30 because it contains little fines. In conclusion, both
types of sand with a hardness of Mohs-7 can be separated using a 5-bar pressure, without risk of crushing. Quartz #3

FIG. 8
Particle size distribution
obtained at different
primary air pressures for
air-dried specimens of
(A) Silica #20-30 and (B)
Quartz #3 sands.
(specimen size = 10 g
and 30 g, respectively).
Note: 1 bar = 100 kPa

particles cannot be well separated at air pressure below 2 bar (200 kPa). Therefore, a primary air pressure of 3 bar
(300 kPa) was used in all later tests in order to ensure consistent results while extending the lifetime of the instrument
that is negatively affected by high pressure (Mootz 2019, personal cummunications).

EFFECT OF SAMPLE MOISTURE CONTENT

VIBRI and RODOS are designed for use with dry specimens. However, sands are not always 100 % dry after ex-
posure to the lab environment for some time. When sands become wet, they may adhere to each other,
as shown in figure 9A, where overlapping particles of Silica #20-30 are shown. Overlapping particles are interpreted
by the software as a single particle having a larger diameter, thus skewing the grain size distribution curve. This may
be overcome with manual deletion of overlapping particles, which is time consuming, or increasing the primary air
pressure, which is not always effective. For quartz specimens, the sticky particles blocked the RODOS tube, as
shown in figure 9B, which halted the test. Water acts as an antilubricant to quartz and silica, thus increasing
the coefficient of static friction, with the antilubricating effect of water typically diminishing with surface roughness
(Horn and Deere 1962). Although both silica and quartz are mineralogically similar, being massive crystal struc-
tures, water appeared to have played a bigger role in increasing the friction of quartz. It appears that the effect of
quartz having lower median S = 0.72 versus 0.8 as well as lower median AR = 0.63 versus 0.8, trumped the effect of
it having a higher median Cx = 0.9 versus 0.97, for quartz and silica, respectively.
To evaluate the effect of moisture content on DIA measurements, Quartz #3 and Silica #20-30 were selected
for tests at different moisture contents. Sands were tested air dry, oven dry, and at moisture content of 0.5 %, 1 %,
2 %, 4 %, and 6 %. Air-dry sand was collected directly from the lab; all other specimens were first oven dried for
24 h according to ASTM D2216-10, Standard Test Methods for Laboratory Determination of Water (Moisture)
Content of Soil and Rock by Mass. Fifty grams of each sand were mixed for each test. After oven drying, a specified
amount of water was thoroughly and evenly mixed with the specimen. The mixing time was limited to less than
1 min to prevent evaporation. After mixing, moist sand was split into two moisture cans: one was used for
immediately conducting DIA, and the other was used to obtain the actual moisture content. Although the tests
were done shortly after sample preparation, moisture evaporation occurred. A comparison of the target and actual
moisture contents is presented in Table 4 for all tests.
The measured particle size distribution of silica and quartz sands at the various moisture contents are
shown in figure 9C and 9D, respectively. The distributions are highly consistent, except for Quartz #3 tested
at a moisture content of 6 %, which is the highest water content percentage tested in this study. It can be noted
that moisture does not adversely affect the DIA imaging system. Therefore, it is suggested that oven drying is not
necessary and that all sands can be tested air dry, as long as the moisture content is less than 3 %.

FIG. 9 Particle size distribution results of sand having different water contents. (A) Overlapping Silica #20-30 particle
images; (B) Sticky Quartz #3 particles blocked the tube; (C) Silica #20-30; and (D) Quartz #3. Tested at: Silica #20-
30 10 g, Quartz #3 30 g, primary air pressure 3 bar (300 kPa). Note: wc represents water content in (C) and (D).

TABLE 4
Comparison of the prepared and actual moisture content

Actual Water Content

Prepared Moisture Content Silica #20-30 Quartz #3

0.50 % 0.42 % 0.41 %

1% 0.90 % 0.94 %
2% 1.74 % 1.82 %
4% 3.70 % 3.86 %
6% 5.20 % 5.43%

EFFECT OF SAMPLE WEIGHT ON MEASUREMENT REPEATABILITY

The accuracy of DIA measurements is closely related to the number of particle images captured. For specimens
having the same weight, sands having a smaller size contain a larger number of particles than larger ones.
Similarly, poorly graded sands have a consistency in particle size, thus a smaller number of particles can
adequately represent the particle size distribution. A low particle count may lead to large statistical errors.
ISO 13322-2, Particle Size Analysis – Image Analysis Methods – Part 2: Dynamic Image Analysis Methods, pro-
vides a method for estimation of the minimum number of particle images required to achieve a mass median
diameter, within 5 % error, with 95 % confidence. The number of particle images varies from a few hundred (for

narrow size distributions) to ten million (for very well-graded particles distributions). However, (1) it is imprac-
tical to count the number of individual tested particles in advance. Additionally, (2) DIA is only able to capture
images of a small percentage of the particles in the specimen. Alternatively, ISO 13320, Particle Size Analysis –
Laser Diffraction Methods, states the test precision can be expressed in terms of the coefficient of variation. In
particular, ISO 13320 recommends that tests should be repeated at least 3 times, on each material and that the
coefficient of variation should be less than 3 % at D50 and 5 % at D10 and D90. ISO 13320 was therefore adopted as
a reference for estimating the effect of sample weight on test accuracy.
The effect of sample weight on particle size distribution was evaluated for four sands: Silica #20-30, Silica
#70-100, Quartz #3, and Quartz # 12 , including two particle shapes and different particle sizes. The uniformity
coefficients, Cu, are 1.23, 1.40, 1.64, and 1.62, corresponding to D50 of 804, 287, 1,155, and 194 μm, respectively.
The change in the value of the coefficient of variation (COV) of D10, D30, D50, and D90 with increasing specimen
weights is depicted in figure 10. Each test was repeated five times. The corresponding number of tested particles is
also shown in figure 10. As expected, the coefficient of variation reduced with increasing specimen weight for all
tests, however, a different amount of sand is required to achieve the same COV for each sand. COV was computed
as the ratio of the standard deviation to the mean value. Quartz #3 requires the largest specimen weight to reach
the reference values of coefficients of variation because of its broad grain size distribution. As a result, a large
number of particles are required to represent the particle size distribution of the whole sample. Silica #20-30 has
the lowest Cu and particles are relatively uniform compared with Quartz #3 and Silica #70-100, with tests with 1 g
of Silica #20-30 are highly consistent.
Our results indicate that COV is inversely proportional to specimen weight. However, the ideal sample weight
is not always available in all circumstances. The COV also depends on the grain size, particle shape, and uniformity
coefficient of the tested sand. An approximate relationship between the required specimen weight, W (g), and
D50 (mm) can be roughly approximated as W = 20 * (D50)3. This relationship yields specimen sizes ranging from
0.1 to 30 g for the 4 tested sands. Nevertheless, a minimum specimen weight of 30 g (approximately 1 oz) is
suggested for all sands in order to achieve a COV of 3 % at D50 and 5 % at D10 and D90 in most sands. The suggested
amount is selected to balance a variety of accuracy, testing, uncertainty, and practicality requirements.

EFFECT OF SPECIMEN SIZE ON TEST DURATION

The effect of sample weight on test duration and the percentage of particles tested was investigated for the four
previously used sands: Silica #20-30, Silica #70-100, Quartz #3, and Quartz # 12 . The weight of the sampled particles
was obtained by first computing the volume of all captured particles in each test, assuming that all particles were
spherical (i.e., using EQPC of each captured particle). Volume was converted to weight, assuming the specific grav-
ity, Gs = 2.65, and the void ratio, e = 0. Next, the measured weight was normalized by the specimen weight. In our
tests, we found that the percentage of particles tested ranged between 3.1 % and 4 % of the specimen weight.
The test duration is related to the specimen weight, particle size, and the feed rate. Optical concentration is a
parameter used to characterize the feed rate during the test. It is defined by ISO 13320 as the fraction of incident
light that is attenuated due to extinction (scattering, absorption, or both) by particles. The feed rate was adjusted
by trial and error during the test to achieve optical concentrations below 0.5 % in order to ensure good particle
separation in all frames. The test durations for quantifying the size and shape of 10 g of the 4 types of sand
investigated were approximately 200, 600, 300, and 800 s for Silica #20-30, Silica #70-100, Quartz #3, and
Quartz # 12 , respectively. These test durations correspond to 20–80 seconds per gram.

Repeatability of DIA Test Results

To quantify the reliability of DIA measurements, both Silica #20-30 and Quartz #3 were each tested 20 times. We
attempted to duplicate the testing environment by cleaning the RODOS unit before each test, which was accom-
plished using high-pressure air to blow away any impurities left from prior tests. All tests were conducted on air-dry
specimens at a primary air pressure of 3 bar (300 kPa). For each test, 10 g of Silica #20-30 and 30 g of Quartz #3 were

FIG. 10 Effect of sample size on measured coefficient of variation of D10, D30, D50, and D90 for (A) Silica #20-30;
(B) Silica #70-100; (C) Quartz #3; (D) Quartz # 21 . All tests conducted with air-dried specimens at a primary air
pressure of 3 bar (300 kPa).

used. A smaller amount of silica was used in order to demonstrate that, for uniform sand, significantly smaller
specimens are required to achieve repeatable results. The mean, standard deviation, and COV of D10, D30, D50, and
D90 for all trials are shown in Table 5. The values of all COVs are less than 1 % for Silica #20-30 and less than 3 %
for Quartz #3. These values are less than the maximum thresholds set by ISO 13320. The results demonstrate the
utility of DIA and that reliable results can be achieved using smaller specimens than the 50–200 g required by
ASTM D6913/D6913M for conventional sieve analysis. Additionally, the data confirm that a smaller specimen
size is required for rounded materials and that 30 g is sufficient for angular materials.

Engineering Performance of DIA

Sixteen different sands were selected to assess the performance of DIA in characterizing a variety of sands
for geotechnical purposes (Table 6). Eight were rounded Silica sands (#12-20, #20-30, #35-50, #40-70, #50-80,

TABLE 5
Repeatability of DIA for 2 sands, Silica #20-30 (10 g) and Quartz #3 (30 g)

Silica #20-30 Quartz #3

Trial D10, μm D30, μm D50, μm D90, μm D10, μm D30, μm D50, μm D90, μm

1 704.86 799.82 846.00 938.37 760.70 1,009.00 1,163.60 1,550.60

2 711.48 803.85 847.56 935.00 747.08 1,009.50 1,159.28 1,555.00
3 703.18 798.15 844.13 936.09 741.22 1,004.68 1,155.47 1,550.45
4 714.17 805.88 850.24 938.97 728.39 991.88 1,135.84 1,524.99
5 705.44 799.98 868.71 937.45 753.26 1,004.23 1,153.98 1,565.86
6 708.43 802.27 847.46 937.85 757.27 1,012.11 1,173.86 1,564.03
7 714.13 805.33 848.96 936.21 684.26 973.34 1,120.05 1,451.74
8 713.47 805.65 850.73 940.91 701.82 997.32 1,141.55 1,499.15
9 707.76 801.67 846.97 937.58 751.92 1,012.93 1,165.57 1,547.89
10 715.57 806.48 850.74 939.27 713.75 989.08 1,130.39 1,465.02
11 707.32 801.74 847.10 937.82 729.81 1,000.92 1,146.67 1,521.31
12 710.53 803.49 870.51 937.54 703.43 999.78 1,164.73 1,498.70
13 707.89 801.43 846.04 935.28 718.27 1,002.28 1,156.97 1,535.22
14 715.75 806.51 849.91 936.73 724.77 999.08 1,144.54 1,544.21
15 708.03 801.8 847.04 937.52 720.41 991.98 1,127.84 1,541.13
16 711.2 803.82 848.59 938.13 727.49 1,000.80 1,165.33 1,572.13
17 712.44 805.03 849.51 938.48 730.09 994.96 1,159.48 1,459.83
18 717.6 808.13 852.91 942.46 747.39 1,017.11 1,175.75 1,540.45
19 717.31 807.01 850.26 936.74 710.09 987.29 1,133.93 1,464.48
20 717.61 808.06 852.62 941.77 726.25 992.1 1,134.97 1,527.26
Average 711.21 803.81 850.80 938.01 728.88 999.52 1,150.49 1,523.97
Standard Deviation 4.45 2.87 6.82 1.96 20.44 10.36 16.14 37.94
Coefficient of Variation 0.63 0.36 0.80 0.21 2.80 1.04 1.40 2.49

Note: All specimens were air-dried and tested at a primary air pressure of 3 bar (300 kPa).

#70-100, #100-140, and #140-270), and eight were angular Quartz sands (#4, #3, #2, #1, # 12 , # 20 , # 30 , and # 40 ).
All tests were conducted on air-dry specimens at a primary air pressure of 3 bar (300 kPa). Two evaluations were
undertaken; the first involved single particle systems where as-received sands were tested without further process-
ing. The second evaluation involved binary particle systems that were made by mixing of two types of sand in
various weight ratios, comprising different particle sizes and shapes.

SINGLE GRANULAR PARTICLE SYSTEMS

Conventional cumulative grain size distribution curves are plotted along with selected captured images in figure 11 for
the 16 test sands. Density distribution curves are also superimposed on the same figures. Cumulative curves represent
the relative particle amount below a particular size, whereas the density curves are frequency (histogram) distributions
of that particle size. Sand particle sizes were described using the EQPC diameter, and the distributions were computed
using volume rather than number distribution. DIA analysis can provide data for every imaged particle, including the
particle area and its longest and shortest dimensions, which are used to compute a variety of shape factors and effective
diameters. D50 and D10 of each of the 16 test sands are summarized in Table 6. The statistics reported include effective
diameters at the extremities, such as D99 and D1; however, these values are not as reliable, unless a very large quantity
of sand is tested because, by definition, each represents approximately 1 % of the volume/mass of the available par-
ticles. The corresponding particle size D1 for all materials tested was 7 μm to 824 μm, whereas D99 ranged from 157 μm
to 3,938 μm (Table 6).
DIA also provides particle shape information, including S, Cx, and AR, for all tested sands. The mean,
median, and standard deviation of these shape descriptors for each sand are summarized in Table 6. Shape

TABLE 6
Particle size and shape information for 16 sands

Sphericity Convexity AR

Standard Standard Standard

Sand D99, μm D50, μm D10, μm D1, μm Mean Median Deviation Mean Median Deviation Mean Median Deviation

Silica #12-20 3,938 1,514 1,040 824 0.79 0.80 0.09 0.97 0.98 0.02 0.78 0.80 0.10
Silica #20-30 1,229 804 724 652 0.78 0.80 0.09 0.96 0.97 0.02 0.79 0.80 0.09
Silica #35-50 907 557 446 353 0.78 0.80 0.10 0.95 0.96 0.03 0.75 0.76 0.10
Silica #40-70 779 533 344 252 0.77 0.79 0.10 0.94 0.95 0.03 0.74 0.75 0.10
Silica #50-80 586 372 261 189 0.77 0.79 0.10 0.93 0.94 0.04 0.74 0.75 0.10
Silica #70-100 429 287 216 153 0.76 0.79 0.11 0.92 0.93 0.04 0.73 0.74 0.11
Silica #100-140 477 200 134 79 0.75 0.78 0.12 0.89 0.90 0.06 0.73 0.74 0.11
Silica #140-270 163 97 63 29 0.75 0.78 0.13 0.84 0.86 0.09 0.73 0.74 0.11
Quartz #4 3,982 2,179 1,252 257 0.68 0.70 0.10 0.90 0.92 0.06 0.60 0.61 0.15
Quartz #3 3,734 1,156 861 138 0.70 0.72 0.10 0.89 0.90 0.05 0.62 0.63 0.15
Quartz #2 735 400 209 44 0.70 0.72 0.12 0.82 0.83 0.09 0.64 0.66 0.15
Quartz #1 513 320 229 62 0.71 0.73 0.12 0.86 0.88 0.08 0.65 0.66 0.14
Quartz # 12 350 194 129 43 0.73 0.75 0.12 0.84 0.87 0.08 0.68 0.69 0.13
Quartz # 20 333 144 65 17 0.74 0.76 0.12 0.77 0.79 0.10 0.66 0.68 0.14
Quartz # 30 196 71 26 9 … … … … … … … … …
Quartz # 40 157 49 17 7 … … … … … … … … …

Note: Particle image resolution is not enough when particles are smaller than 40 μm.

descriptors were available for 14 of the 16 sands. Particle image resolution was insufficient to compute the shape
descriptors of Quartz # 30 and # 40 . The current version of DIA fails in determining particle shapes when particles
are smaller than 40 μm. For all tested sands, the mean and median values of Cx are larger than their S and AR
counterparts, whereas the values of standard deviation of Cx are noticeably smaller than S and AR. Thus, Cx is
the least sensitive parameter for capturing the variation of particle shape morphology compared with S and AR.
Cx is believed to better capture local features, such as surface morphology, whereas S and AR are better descriptors
of the overall particle shape. Values of S and AR were close for silica sand because of its rounded nature
(S/AR = 0.98–1.04). As for the quartz sands, values of S are consistently larger than AR, (S/AR = 1.08 to
1.14). Values of Cx were larger for silica sand than for quartz sand (0.93 ± 0.04 versus 0.84 ± 0.06), suggesting
that quartz has a rougher surface morphology.
In silica sand, the values of S and AR increase with the increase in particle size but generally decrease for
quartz sand. Thus, the larger the particle of silica sand, the rounder it is, whereas large particles of quartz tend to
be more angular. This trend is easier to observe for silica sand in SEM images than for quartz sand (fig. 1), which
further highlights the utility of DIA. Variation in AR for silica and quartz sands was limited, with the standard
deviation in AR being 0.10 ± 0.01 and 0.13 ± 0.02, respectively, with quartz exhibiting a somewhat larger variation
in its morphology, as measured by the change in standard deviation of AR, within each size. Variation in S and Cx
decreased with increasing particle size in silica sand but generally decreased for quartz sand.
These results cumulatively point to the fact that silica sands lose their roundness and uniformity, at the same
time, as they decrease in size. In contrast, smaller quartz sand particles are rounder than their larger counterparts.
The data also point out that quartz is somewhat rougher than silica, in most, but not all sizes. These inferences
would be difficult to identify by visual observation of SEM images. DIA quantified a large number of particle
shape information and described the roundness and angularity of both sands accurately.

BINARY PARTICLE SYSTEMS

Analysis of many geotechnical situations involves the study of materials having wide grain size distributions and at
times, binary grain systems or gap-graded systems. For example, binary particle systems occur whenever particle

FIG. 11 Particle size, density distribution, and representative captured particle images of 16 sands. (A) Silica #12-20,
#20-30, #35-50, #40-70; (B) Silica #50-80, #70-100, #100-140, #140-270; (C) Quartz #4, #3, #2, #1;
(D) Quartz # 21 , # 02 , # 03 , # 04 . All specimens were 30 g, air-dried, and tested at a primary air pressure of 3 bar
(300 kPa).

crushing takes place (e.g., pile driving) or when internal instability of soils leads to consequential results. Four
granular mixture families were employed to investigate the capability of DIA for identifying the size distribution
within soils having widely varying particle size or shape distributions. Each family was made of 2 sands that were
mixed such that the mass fractions of both sands during each test was varied from 1:9 to 9:1, resulting in 9 mixtures
within each family. Five sands having different size and shape were employed to make 36 distinct binary mixtures.
Particle size distribution curves of all mixtures are plotted in figure 12. The mix ratios as well as the
resulting particle size of D10, D30, D50, D60, and D90 are presented in Table 7. Figure 12A represents mixtures
of rounded particles (Silica #12-20 and #140-270), and figure 12B represents mixtures of angular particles
(Quartz #3 and # 12 ). Figure 12B and 12D represents mixtures of round and angular particles where large
rounded particles are mixed with fine angular sand (Silica #20-30 and Quartz # 12 ) in figure 12B, and
figure 12D represents large angular particles mixed with fine rounded sand (Quartz #3 and Silica #140-270).
All distributions are therefore gap graded. The uniformity coefficients, Cu, were 1.64–15.17, 1.31–5.28,
1.75–6.86, 1.72–14.15, whereas the coefficients of curvature, Cc, were 0.40–3.25, 0.32–3.28, 0.14–6.98, and
0.14–7.52 for the 4 families (Table 7).

FIG. 12 Ability of DIA to characterize the grain size distribution of gap graded soils. (A) Mixture of Silica #12-20 and
#140-270, (B) Mixture of Silica #20-30 and Quartz # 21 , (C) Mixture of Quartz #3 and # 21 , (D) Mixture of Quartz #3
and Silica #140-270.

In all 36 mixtures, the maximum value of Cu typically occurs when materials are mixed 5:5 ± 1:1.
For example, Cu = 14.15 and 15.17 occurred, respectively, at mix ratios of 5:5 and 6:4 in the Quartz #3
and Silica #140-270 and Silica #12-20 and #140-270 families. Values of D10 gradually increased with the
increase in the weight fraction of the coarser portion in each mixture. The values of D50 and D60 are typically
dominated by the larger of the 2 mixed sands when the fine fraction is less than 70 % of the specimen
weight for all mix families. Cu typically decreases as the mix becomes dominated by one size (9:1 or 1:9).
Values of Cc typically showed an opposite trend to that of Cu, with the minimum values occurring at specimen
weight ratio of 4:6 and 5:5 in all families. The lowest value of Cc ranged from 0.14 to 0.4 in all 4 families
(Table 7).
For mixtures made of nonoverlapping grain size distributions, D50 Mix should be D5 Coarse > D50 Mix > D95 Fine.
It is, however, concerning that the values of D50 of the 5:5 mixtures are larger than the D5 of the coarse fraction
except for Quartz #3 and # 12 (Table 8). This occurs because DIA captures many but not all the particles, and
D50 is influenced by presence of larger particles than smaller ones. In particular, because the distribution is
influenced by area/weight, the presence of a single larger particle can affect D50 more than presence of a small
one. When a small overlap in the two parent grain size distributions occur, D50 should be D50 Mix < D95 Fine,

TABLE 7
Particle size, uniformity coefficient, and coefficient of gradation for four granular mixture families

Silica #12-20 and #140-270

Weight Ratio #12-20 (g) #140-270 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

1:9 5 45 1,420 114 105 87 65 1.75 1.02

2:8 10 40 1,440 127 111 91 67 1.90 0.97
3:7 15 35 1,639 1,041 154 102 73 14.23 0.14
4:6 20 30 1,682 1,136 1,000 111 77 14.72 0.14
5:5 25 25 2,084 1,203 1,074 136 82 14.6 0.19
6:4 30 20 2,306 1,388 1,211 971 92 15.17 7.38
7:3 35 15 2,305 1,384 1,252 1,025 101 13.67 7.52
8:2 40 10 2,334 1,450 1,344 1,116 140 10.39 6.14
9:1 45 5 2,703 1,511 1,425 1,164 920 1.64 0.97

Silica #20-30 and Quartz # 12

#20-30 (g) # 12 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

Weight Ratio

1:9 5 45 773 229 210 178 141 1.63 0.98

2:8 10 40 875 256 228 186 145 1.76 0.93
3:7 15 35 898 296 249 195 149 1.98 0.86
4:6 20 30 919 747 328 216 158 4.74 0.40
5:5 25 25 921 807 741 227 153 5.28 0.42
6:4 30 20 927 828 796 288 165 5.02 0.61
7:3 35 15 929 842 813 704 181 4.65 3.25
8:2 40 10 935 857 831 771 217 3.95 3.20
9:1 45 5 934 860 835 786 657 1.31 1.09

Quartz #3 and Quartz # 12

#3 (g) # 12 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

Weight Ratio

1:9 5 45 766 219 199 168 125 1.75 1.03

2:8 10 40 1,204 240 215 176 131 1.83 0.99
3:7 15 35 1,292 284 236 184 137 2.07 0.87
4:6 20 30 1,379 827 280 196 141 5.85 0.33
5:5 25 25 1,382 945 653 211 146 6.46 0.32
6:4 30 20 1,454 1,089 978 266 159 6.86 0.41
7:3 35 15 1,403 1,122 1,025 522 169 6.66 1.44
8:2 40 10 1,526 1,185 1,091 860 196 6.05 3.18
9:1 45 5 1,520 1,171 1,092 928 262 4.47 2.81

Quartz #3 and Silica #140-270

Weight Ratio #3 (g) #140-270 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

1:9 5 45 1,201 116 106 89 67 1.72 1.02

2:8 10 40 1,300 127 111 91 68 1.85 0.96
3:7 15 35 1,442 822 132 100 74 11.18 0.16
4:6 20 30 1,507 1,032 802 107 77 13.45 0.14
5:5 25 25 1,534 1,148 1,004 124 81 14.15 0.17
6:4 30 20 1,566 1,193 1,082 653 87 13.67 4.11
7:3 35 15 1,555 1,225 1,140 925 100 12.3 6.98
8:2 40 10 1,556 1,242 1,166 981 124 10.05 6.25
9:1 45 5 1,603 1,254 1,174 1,008 662 1.89 1.22

TABLE 8
Comparison of particle size D50 for four granular mixture families, at select weight ratios

Silica #12-20 and #140-270

Mix Coarse Fraction #12-20 Fine Fraction #140-270

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 1,000 3,612 1,514 973 147 97 52

5:5 1,074
6:4 1,211

Silica #20-30 and Quartz # 12

Mix Coarse Fraction #20-30 Fine Fraction # 12

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 328 1,029 804 684 299 194 103

5:5 741
6:4 796

Quartz #3 and Quartz # 12

Mix Coarse Fraction #3 Fine Fraction # 12

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 280 2,594 1,156 682 299 194 103

5:5 653
6:4 978

Quartz #3 and Silica #140-270

Mix Coarse Fraction #3 Fine Fraction #140-270

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 802 2,594 1,156 682 147 97 52

5:5 1,004
6:4 1,082

but the results of D50 are larger than D95 (Quartz #3 and # 12 ), confirming that DIA tends to capture a larger weight of
coarser particles than finer ones.

Limitation
DIA offers many advantages, but it also has a number of limitations, notably, the following:

• Overlapping particles are regarded as one particle by the software, which could influence the accuracy of test
results of larger particle diameters, such as D99. This is not significant for routine work but may potentially be
addressed in the future by incorporating automatic identification of particles and their overlaps. Segmentation
techniques such as watershed algorithm may then be employed to separate overlapping particles.
• Like many soil tests, it is difficult to accommodate a variety of sizes in the DIA apparatus. The maximum
particle size measured by DIA is approximately 4 mm. However, DIA and sieve analysis maybe combined
together to study well-graded soils with particle sizes larger than 4 mm.
• Particle roundness, which, at this time, is not available in the current version of DIA, is an important parameter
for characterizing particle shapes of sand. The images provided by DIA can however be employed together
with methods published by Zheng and Hryciw (2015, 2016b) to compute particle roundness automatically.

Conclusion
DIA involves transporting and separating granular particles from each other using an air stream and imaging the
sparse particles using a high-frame-rate camera. The method proved to be feasible, repeatable, and accurate for
providing particle size distributions. DIA offers a number of advantages in comparison with conventional sieve
analysis, laser diffraction, or SEM for measuring the physical properties of sand. In particular, the method is
quick, requires small specimen sizes, provides quantitative information on a large number of the particles,
and is capable of measuring particles size as large as 3,938 μm and down to 7 μm. Particles’ morphology is
analyzed using both size and shape descriptors. Particle size distributions can be computed using a number
of size descriptors, including EQPC or Feret diameters. Feret-min was shown to be the closest distribution
to that obtained from sieve analysis. A variety of particle size descriptors are provided, including D10, D30,
D50, D60, and D90. Two-dimensional shape descriptors including AR, S, and Cx are provided by DIA for each
imaged particle, provided that the particle is larger than 40 μm. The voluminous particle size and shape infor-
mation provided is helpful in understanding particle shape distribution within the specimen.
The effects of primary air pressure, moisture content, and specimen weight on the accuracy of DIA mea-
surements was investigated. A primary air pressure of 3 bar (300 kPa) was found sufficient to separate all particles
without posing a risk of particle crushing. Specimen moisture content should be less than 3 % to ensure that
particles do not adhere to each other. The repeatability of DIA increases as the number of imaged particles in-
crease and is generally inversely proportional to the specimen weight. However, grain size, particle shape, and the
uniformity coefficient of the specimen combine to determine the necessary number of particle images needed to
achieve repeatable DIA test results. A minimum specimen weight of 30 g (approximately 1 oz) is suggested for all
sands in order to achieve a COV of 3 % at D50 and 5 % at D10 and D90 in most sands.
Particle size and shape distributions were obtained using DIA for 52 granular systems. Grain size distribution
curves were obtained for 16 different sands comprising both rounded and angular shapes and a variety of sizes
ranging from D50 = 49 to 2,179 μm. The distribution of 2-D shape descriptors (S, Cx, AR) was obtained for each
sand. Cx was found to be the least sensitive parameter for capturing the variation of particle shape morphology,
compared with S and AR. Additionally, 36 binary particle systems belonging to 4 gap-graded sand mixture fam-
ilies were tested. The results demonstrate that DIA was able to identify particle size distribution and shape in-
formation for granular systems comprising a wide range of particle shapes and sizes.
Therefore, this study demonstrates that DIA can potentially replace standard sieve analysis in routine geo-
technical engineering. This overcomes the limitation of standard sieve analysis for particles finer than 250 μm.

ACKNOWLEDGMENTS
A Sympatec QicPic DIA device was employed for all tests. All sands were commercially sourced from AGSCO
Corp. The technical assistance and of Dr. Kay Mootz of Sympatec is gratefully acknowledged.

References
Alshibli, K. A., A. M. Druckrey, R. I. Al-Raoush, T. Weiskittel, and N. V. Lavrik. 2015. “Quantifying Morphology of Sands
Using 3D Imaging.” Journal of Materials in Civil Engineering 27, no. 10 (October): 04014275. https://doi.org/10.1061/(asce)
mt.1943-5533.0001246
Altuhafi, F. N. and M. R. Coop. 2011. “Changes to Particle Characteristics Associated with the Compression of Sands.”
Géotechnique 61, no. 6 (June): 459–471. https://doi.org/10.1680/geot.9.P.114
Altuhafi, F., C. O’Sullivan, and I. Cavarretta. 2013. “Analysis of an Image-Based Method to Quantify the Size and Shape of Sand
Particles.” Journal of Geotechnical and Geoenvironmental Engineering 139, no. 8 (August): 1290–1307. https://doi.org/10.
1061/(asce)gt.1943-5606.0000855
Altuhafi, F. N., M. R. Coop, and V. N. Georgiannou. 2016. “Effect of Particle Shape on the Mechanical Behavior of Natural
Sands.” Journal of Geotechnical and Geoenvironmental Engineering 142, no. 12 (December): 04016071. https://doi.org/10.
1061/(asce)gt.1943-5606.0001569
Association Francaise de Normalisation. 1990. Test Sieves - Metal Wire Cloth, Perforated Metal Plate and Electroformed Sheet -
Nominal Sizes of Openings. NF ISO 565:1990. Paris: Association Francaise de Normalisation, 1990.

ASTM International. 2010. Standard Test Methods for Laboratory Determination of Water (Moisture) Content of Soil and Rock
by Mass. ASTM D2216-10. West Conshohocken, PA: ASTM International, approved July 1, 2010. https://doi.org/10.1520/
D2216-10
ASTM International. 2016. Standard Practice for Characterization of Particles. ASTM F1877-16. West Conshohocken, PA:
ASTM International, approved November 2019. https://doi.org/10.1520/F1877-16
ASTM International. 2017. Standard Test Methods for Particle-Size Distribution (Gradation) of Soils Using Sieve Analysis.
ASTM D6913/D6913M-17. West Conshohocken, PA: ASTM International, approved April 15, 2017. https://doi.org/10.
1520/D6913_D6913M-17
ASTM International. 2017. Standard Specification for Woven Wire Test Sieve Cloth and Test Sieves. ASTM E11-17. West
Conshohocken, PA: ASTM International, approved April 1, 2017. https://doi.org/10.1520/E0011-17
Bareither, C. A., T. B. Edil, C. H. Benson, and D. M. Mickelson. 2008. “Geological and Physical Factors Affecting the Friction
Angle of Compacted Sands.” Journal of Geotechnical and Geoenvironmental Engineering 134, no. 10 (October): 1476–1489.
https://doi.org/10.1061/(asce)1090-0241(2008)134:10(1476)
British Standards Institute. 2016. Methods of Test for Soils for Civil Engineering Purposes. General Requirements and Sample
Preparation. BS 1377-1:2016. London: British Standards Institute, 2016.
Cepuritis, R., E. J. Garboczi, S. Jacobsen, and K. A. Snyder. 2017. “Comparison of 2-D and 3-D Shape Analysis of Concrete
Aggregate Fines from VSI Crushing.” Powder Technology 309 (March): 110–125. https://doi.org/10.1016/j.powtec.2016.12.037
China National Standards. 2007. Test Methods of Soils for Highway Engineering. JTG E40-2007. Shenzen, China: China
National Standards, 2007.
Cho, G. C., J. Dodds, and J. C. Santamarina. 2006. “Particle Shape Effects on Packing Density, Stiffness, and Strength: Natural
and Crushed Sands.” Journal of Geotechnical and Geoenvironmental Engineering 132, no. 5 (November): 591–602. https://
doi.org/10.1061/(asce)1090-0241(2006)132:5(591)
German Institute for Standardization. 2002. Soil Quality - Determination of Particle Size Distribution in Mineral Soil Material -
Method by Sieving and Sedimentation. DIN ISO 11277:2002. Berlin: German Institute for Standardization, 2002.
Horn, H. M. and D. U. Deere. 1962. “Frictional Characteristics of Minerals.” Géotechnique 12, no. 4 (May): 319–335. https://doi.
org/10.1680/geot.1962.12.4.319
Howard, A. K. 1984. “The Revised ASTM Standard on the Unified Classification System.” Geotechnical Testing Journal 7, no. 4
(December): 216–222. https://doi.org/10.1520/gtj10505j
International Organization for Standardization. 2006. Particle Size Analysis – Image Analysis Methods – Part 2: Dynamic Image
Analysis Methods. ISO 13222-2:2006. Geneva, Switzerland: International Organization for Standardization, 2006.
International Organization for Standardization. 2006. Test Sieves – Technical Requirements and Testing – Part 1: Test Sieves of
Metal Wire Cloth. ISO 3310-1:2006. Geneva, Switzerland: International Organization for Standardization, 2006.
International Organization for Standardization. 2008. Representation of Results of Particle Size Analysis – Part 6: Descriptive and
Quantitative Representation of Particle Shape and Morphology. ISO 9276-6:2008. Geneva, Switzerland: International
Organization for Standardization, 2008.
International Organization for Standardization. 2009. Particle Size Analysis – Laser Diffraction Methods. ISO 13320:2009.
Geneva, Switzerland: International Organization for Standardization, 2009.
International Organization for Standardization. 2009. Soil Quality – Determination of Particle Size Distribution in Mineral Soil
Material – Method by Sieving and Sedimentation. ISO 11277:2009. Geneva, Switzerland: International Organization for
Standardization, 2009.
Janoo, V. 1998. Quantification of Shape, Angularity, and Surface Texture of Base Course Materials, No. CRREL-SR-98-1.
Hanover, NH: Cold Regions Research and Engineering Lab.
Japanese Standards Association. 2006. Test Sieves Part 1: Test Sieves of Metal Wire Cloth. JIS Z 8801-1:2006. Tokyo: Japanese
Standards Association, 2006.
Japanese Standards Association. 2009. Test Method for Particle Size Distribution of Soils. JIS A 1204:2009. Tokyo: Japanese
Standards Association, 2009.
Krumbein, W. C. 1941. “Measurement and Geological Significance of Shape and Roundness of Sedimentary Particles.” Journal
of Sedimentary Research 11, no. 2 (August): 64–72. https://doi.org/10.1306/d42690f3-2b26-11d7-8648000102c1865d
Krumbein, W. C. and L. L. Sloss. 1951. Stratigraphy and Sedimentation. San Francisco, CA: W. H. Freeman and Company.
Kuo, C.-Y. and R. Freeman. 2000. “Imaging Indices for Quantification of Shape, Angularity, and Surface Texture
of Aggregates.” Transportation Research Record: Journal of the Transportation Research Board 1721, no. 1 (January):
57–65. https://doi.org/10.3141/1721-07
McDowell, G. R., M. D. Bolton, and D. Robertson. 1996. “The Fractal Crushing of Granular Materials.” Journal of the Mechanics
and Physics of Solids 44, no. 12 (December): 2079–2101. https://doi.org/10.1016/S0022-5096(96)00058-0
Mora, C. F. and A. K. H. Kwan. 2000. “Sphericity, Shape Factor, and Convexity Measurement of Coarse Aggregate for Concrete
Using Digital Image Processing.” Cement and Concrete Research 30, no. 3 (March): 351–358. https://doi.org/10.1016/
s0008-8846(99)00259-8
Nalluri, V. R., P. Schirg, X. Gao, A. Virdis, G. Imanidis, and M. Kuentz. 2010. “Different Modes of Dynamic Image Analysis in
Monitoring of Pharmaceutical Dry Milling Process.” International Journal of Pharmaceutics 391, nos. 1–2 (May): 107–114.
https://doi.org/10.1016/j.ijpharm.2010.02.027
Pentland, A. 1927. “A Method of Measuring the Angularity of Sands.” Proceedings and Transactions of the Royal Society
of Canada 21, no. 3: 93.

Polakowski, C., A. Sochan, A. Bieganowski, M. Ryżak, R. Földényi, and J. Tóth. 2014. “Influence of the Sand Particle Shape on
Particle Size Distribution Measured by Laser Diffraction Method.” International Agrophysics 28, no. 2 (April): 195–200.
https://doi.org/10.2478/intag-2014-0008
Powers, M. C. 1953. “A New Roundness Scale for Sedimentary Particles.” Journal of Sedimentary Research 23, no. 2 (June):
117–119. https://doi.org/10.1306/d4269567-2b26-11d7-8648000102c1865d
Riley, N. A. 1941. “Projection Sphericity.” Journal of Sedimentary Research 11, no. 2 (August): 94–95. https://doi.org/10.1306/
D426910C-2B26-11D7-8648000102C1865D
Rodriguez, J., T. Edeskär, and S. Knutsson. 2013. “Particle Shape Quantities and Measurement Techniques: A Review.”
The Electronic Journal of Geotechnical Engineering 18: 169–198.
Rousé, P. C., R. J. Fannin, and D. A. Shuttle. 2008. “Influence of Roundness on the Void Ratio and Strength of Uniform Sand.”
Géotechnique 58, no. 3 (April): 227–231. https://doi.org/10.1680/geot.2008.58.3.227
Shin, H. and J. C. Santamarina. 2013. “Role of Particle Angularity on the Mechanical Behavior of Granular Mixtures.” Journal of
Geotechnical and Geoenvironmental Engineering 139, no. 2 (February): 353–355. https://doi.org/10.1061/(asce)gt.
1943-5606.0000768
Turcotte, D. L. 1986. “Fractals and Fragmentation.” Journal of Geophysical Research: Solid Earth 91, no. B2 (February):
1921–1926. https://doi.org/10.1029/JB091iB02p01921
Wadell, H. 1932. “Volume, Shape, and Roundness of Rock Particles.” The Journal of Geology 40, no. 5 (July/August): 443–451.
https://doi.org/10.1086/623964
Wadell, H. 1933. “Sphericity and Roundness of Rock Particles.” The Journal of Geology 41, no. 3 (April/May): 310–331. https://
doi.org/10.1086/624040
Wadell, H. 1935. “Volume, Shape, and Roundness of Quartz Particles.” The Journal of Geology 43, no. 3 (April/May): 250–280.
https://doi.org/10.1086/624298
White, D. J. 2003. “PSD Measurement Using the Single Particle Optical Sizing (SPOS) Method.” Géotechnique 53, no. 3 (April):
317–326. https://doi.org/10.1680/geot.2003.53.3.317
Yang, J. and X. D. Luo. 2015. “Exploring the Relationship between Critical State and Particle Shape for Granular Materials.”
Journal of the Mechanics and Physics of Solids 84 (November): 196–213. https://doi.org/10.1016/j.jmps.2015.08.001
Yu, W. and B. C. Hancock. 2008. “Evaluation of Dynamic Image Analysis for Characterizing Pharmaceutical Excipient
Particles.” International Journal of Pharmaceutics 361, nos. 1–2 (September): 150–157. https://doi.org/10.1016/j.
ijpharm.2008.05.025
Zheng, J. and R. D. Hryciw. 2015. “Traditional Soil Particle Sphericity, Roundness and Surface Roughness by Computational
Geometry.” Géotechnique 65, no. 6 (June): 494–506. https://doi.org/10.1680/geot.14.P.192
Zheng, J. and R. D. Hryciw. 2016a. “Index Void Ratios of Sands from Their Intrinsic Properties.” Journal of Geotechnical and
Geoenvironmental Engineering 142, no. 12 (December): 06016019. https://doi.org/10.1061/(asce)gt.1943-5606.0001575
Zheng, J. and R. D. Hryciw. 2016b. “Roundness and Sphericity of Soil Particles in Assemblies by Computational Geometry.”
Journal of Computing in Civil Engineering 30, no. 6 (November): 04016021. https://doi.org/10.1061/(ASCE)CP.1943-5487.
0000578

Evaluation of Grain Size Distribution by Digital Image Processing
No ratings yet
Evaluation of Grain Size Distribution by Digital Image Processing
11 pages
Development and Testing of A Portable In-Situ Near-Surface Soil Characterization System
No ratings yet
Development and Testing of A Portable In-Situ Near-Surface Soil Characterization System
362 pages
Breakout Play (Trend Following) - Trading Plan - Full (Sample)
91% (11)
Breakout Play (Trend Following) - Trading Plan - Full (Sample)
15 pages
Hyonsohk 1
No ratings yet
Hyonsohk 1
334 pages
Soil Characterization Using Digital Image Processing: Lehigh Preserve
No ratings yet
Soil Characterization Using Digital Image Processing: Lehigh Preserve
68 pages
Correlations of Soil and Rock Properties in Geotechnical Engineering
95% (19)
Correlations of Soil and Rock Properties in Geotechnical Engineering
236 pages
SI Book References and Standards
No ratings yet
SI Book References and Standards
26 pages
Geotechnics - Automatic Feature Detection and Interpretation in Borehole (HiRAT - Acoutic Televiewer) PHD
No ratings yet
Geotechnics - Automatic Feature Detection and Interpretation in Borehole (HiRAT - Acoutic Televiewer) PHD
130 pages
Main Papr CNN
No ratings yet
Main Papr CNN
31 pages
Full-Scale Field Tests of Different Types of Piles: Jialin Zhou Erwin Oh
No ratings yet
Full-Scale Field Tests of Different Types of Piles: Jialin Zhou Erwin Oh
293 pages
基于人工智能与机器学习方法的颗粒材料休止角三维离散元建模
No ratings yet
基于人工智能与机器学习方法的颗粒材料休止角三维离散元建模
24 pages
Back and Al, 2025
No ratings yet
Back and Al, 2025
21 pages
Micromachines 15 01293
No ratings yet
Micromachines 15 01293
26 pages
Victoria Code of Practice For Using Concrete Pump
0% (1)
Victoria Code of Practice For Using Concrete Pump
56 pages
Modul English PSPK
No ratings yet
Modul English PSPK
139 pages
AgroGeophysics Seminar Abstarct-Booklet
No ratings yet
AgroGeophysics Seminar Abstarct-Booklet
16 pages
1 s2.0 S0026265X18313626 Main
No ratings yet
1 s2.0 S0026265X18313626 Main
9 pages
Paper Avizo PSD
No ratings yet
Paper Avizo PSD
14 pages
Evaluation of A Simple and Affordable
No ratings yet
Evaluation of A Simple and Affordable
14 pages
AE in Crushing of Individual Particles
No ratings yet
AE in Crushing of Individual Particles
11 pages
Particle Roundness and Sphericity From Images of Assemblies by Chart Estimates and Computer Methods
No ratings yet
Particle Roundness and Sphericity From Images of Assemblies by Chart Estimates and Computer Methods
15 pages
2-An Alternative Method For The Particle Size Distributionimage Processing
No ratings yet
2-An Alternative Method For The Particle Size Distributionimage Processing
8 pages
Sedimaging Technique in Soil Classification
No ratings yet
Sedimaging Technique in Soil Classification
35 pages
Sensor Systems For Mapping Soil
No ratings yet
Sensor Systems For Mapping Soil
22 pages
OLOMOLATAN
No ratings yet
OLOMOLATAN
142 pages
Automated Soil Physical Parameter Assessment Using
No ratings yet
Automated Soil Physical Parameter Assessment Using
12 pages
The Use of Artificial Neural Networks in The Determination of Soil Grain Composition
No ratings yet
The Use of Artificial Neural Networks in The Determination of Soil Grain Composition
9 pages
25 Fractal Features of Soil Particle Size Distribution in Newly Formed Wetlands
No ratings yet
25 Fractal Features of Soil Particle Size Distribution in Newly Formed Wetlands
9 pages
PSA Paper Geomorph Treatise 2013 Preprint
No ratings yet
PSA Paper Geomorph Treatise 2013 Preprint
31 pages
1 s2.0 S0032591018308362 Main
No ratings yet
1 s2.0 S0032591018308362 Main
12 pages
Agronomy 12 03063 v2
No ratings yet
Agronomy 12 03063 v2
16 pages
Mapping of Soil Organic Matter (SOM) Using Hyperspectral Imaging and Non-Imaging Data - A Review
No ratings yet
Mapping of Soil Organic Matter (SOM) Using Hyperspectral Imaging and Non-Imaging Data - A Review
3 pages
Quantitative Research On The Evolution of Mesoparameters at Out of The Shear Band of The Air-Dried Clay
No ratings yet
Quantitative Research On The Evolution of Mesoparameters at Out of The Shear Band of The Air-Dried Clay
15 pages
Esfuerzos en Vigas - PDF
No ratings yet
Esfuerzos en Vigas - PDF
6 pages
Ijsdp 18.03 16
No ratings yet
Ijsdp 18.03 16
11 pages
Water Infrustructure
No ratings yet
Water Infrustructure
25 pages
Analysis of Agricultural Soil PH Using Digital Image Processing
No ratings yet
Analysis of Agricultural Soil PH Using Digital Image Processing
5 pages
Estimation of Dielectric Constant of Herbal Based Soil in Relation To Microwave Remote Sensing
No ratings yet
Estimation of Dielectric Constant of Herbal Based Soil in Relation To Microwave Remote Sensing
5 pages
Ijsei 43815 06
No ratings yet
Ijsei 43815 06
6 pages
21-2017-IJMRE-Thomas Bamford-A Real Time Analysis of Post Blast Rock Fragmentation Using UAV Technology
No ratings yet
21-2017-IJMRE-Thomas Bamford-A Real Time Analysis of Post Blast Rock Fragmentation Using UAV Technology
19 pages
CE 344 Geotechnical Engineering Sessional-I (Lab Manual) : January 2012
No ratings yet
CE 344 Geotechnical Engineering Sessional-I (Lab Manual) : January 2012
121 pages
ASI Show Orlando 2025 Exhibitor List
No ratings yet
ASI Show Orlando 2025 Exhibitor List
16 pages
Correlations of Soil and Rock Properties in Geotechnical Engineering岩土工程中岩土特性的相关性
No ratings yet
Correlations of Soil and Rock Properties in Geotechnical Engineering岩土工程中岩土特性的相关性
24 pages
Index: Soil Properties and Their Correlations, Second Edition. Michael Carter and Stephen P. Bentley
No ratings yet
Index: Soil Properties and Their Correlations, Second Edition. Michael Carter and Stephen P. Bentley
6 pages
Evaluating Reflective Spectroscopy For Predicting Soil Properties in Gajapati District, Odisha
No ratings yet
Evaluating Reflective Spectroscopy For Predicting Soil Properties in Gajapati District, Odisha
14 pages
Chemistry Class 10
No ratings yet
Chemistry Class 10
8 pages
E. Journal Type Template
No ratings yet
E. Journal Type Template
12 pages
Problem Solving 11 20
No ratings yet
Problem Solving 11 20
10 pages
Objectives
No ratings yet
Objectives
7 pages
CMT 3
No ratings yet
CMT 3
7 pages
Book Sizes
No ratings yet
Book Sizes
9 pages
Datasheet - Hexaband APE4518R14v06 PDF
No ratings yet
Datasheet - Hexaband APE4518R14v06 PDF
3 pages
Soil Parameter Detection of Soil Test Kit-Treated Soil Samples Through Image Processing With Crop and Fertilizer Recommendation
No ratings yet
Soil Parameter Detection of Soil Test Kit-Treated Soil Samples Through Image Processing With Crop and Fertilizer Recommendation
10 pages
Direct Measurement and Prediction of Bulk Density On Alluvial Soil in Ishiagu
No ratings yet
Direct Measurement and Prediction of Bulk Density On Alluvial Soil in Ishiagu
28 pages
Analysis of Grain Size
No ratings yet
Analysis of Grain Size
12 pages
Food Safety, Sanitation and Hygience
No ratings yet
Food Safety, Sanitation and Hygience
21 pages
Field Tests of Soil Properties Under Water Tanks: The Open Civil Engineering Journal
No ratings yet
Field Tests of Soil Properties Under Water Tanks: The Open Civil Engineering Journal
13 pages
Complexity Analysis of Peat Soil Density PDF
No ratings yet
Complexity Analysis of Peat Soil Density PDF
6 pages
GOL Project
No ratings yet
GOL Project
7 pages
Edraky - SD
No ratings yet
Edraky - SD
29 pages
79-Article Text-432-1-10-20190501
No ratings yet
79-Article Text-432-1-10-20190501
9 pages
University of Okara: Advertisement No. 2/2020
No ratings yet
University of Okara: Advertisement No. 2/2020
3 pages
5 Tran NguyenEdil2011 ThecharacteristicsofPVDSmearzone
No ratings yet
5 Tran NguyenEdil2011 ThecharacteristicsofPVDSmearzone
14 pages
Stabilization of Soil in Road Construction Using Lime and Fly Ash
No ratings yet
Stabilization of Soil in Road Construction Using Lime and Fly Ash
9 pages
Soil Analysis
No ratings yet
Soil Analysis
4 pages
Abstract Fernandez-Ibanezetal.2018 AAPG CTL
No ratings yet
Abstract Fernandez-Ibanezetal.2018 AAPG CTL
2 pages
YLSTD30-40K01小功率直流充电桩用户手册User Manua V1 - (EN&CN) ) 已校对
No ratings yet
YLSTD30-40K01小功率直流充电桩用户手册User Manua V1 - (EN&CN) ) 已校对
17 pages
Lecture 8 - Transport Layer
No ratings yet
Lecture 8 - Transport Layer
50 pages
Lecture 7-2
No ratings yet
Lecture 7-2
37 pages
Grade 7 History Term 1 Worksheets 2023
No ratings yet
Grade 7 History Term 1 Worksheets 2023
23 pages
S71937 - Enabling Intelligent Storage To Process Data For Ai Application Ibm
No ratings yet
S71937 - Enabling Intelligent Storage To Process Data For Ai Application Ibm
21 pages
Cardiosync Corporate Business Plan
No ratings yet
Cardiosync Corporate Business Plan
7 pages
Daylighting Streams Text
No ratings yet
Daylighting Streams Text
6 pages
Current Research Topics in Optical Sensors and Laser Diagnostics
No ratings yet
Current Research Topics in Optical Sensors and Laser Diagnostics
17 pages
NTFK VOL 104 2 (3) 2017 - Henri Rikander - The Use of Electroshock Weapons by The Finnish Police 2016
No ratings yet
NTFK VOL 104 2 (3) 2017 - Henri Rikander - The Use of Electroshock Weapons by The Finnish Police 2016
34 pages
Group 2 - Aspects of Connected Speech
No ratings yet
Group 2 - Aspects of Connected Speech
31 pages
WhatsApp Chat PDF
No ratings yet
WhatsApp Chat PDF
1 page
TOMEI Camshaft Spec Card TOMEI Camshaft Spec Card TOMEI Camshaft Spec Card
No ratings yet
TOMEI Camshaft Spec Card TOMEI Camshaft Spec Card TOMEI Camshaft Spec Card
1 page
Kelvin Bridges - Seaward Support
No ratings yet
Kelvin Bridges - Seaward Support
10 pages
MaheswariVeni Auth Nagercoil
No ratings yet
MaheswariVeni Auth Nagercoil
2 pages
Keberhasilan Media Promosi Judi Online Dalam Menarik Minat Masyarakat-1
No ratings yet
Keberhasilan Media Promosi Judi Online Dalam Menarik Minat Masyarakat-1
10 pages
Fallout 4 Bobblehead and Magazine Guide - Zone 1
No ratings yet
Fallout 4 Bobblehead and Magazine Guide - Zone 1
1 page
HDBS Parameters
No ratings yet
HDBS Parameters
1 page
MODULE 4 MAT Antepartum Flexible Learning
No ratings yet
MODULE 4 MAT Antepartum Flexible Learning
2 pages
Flexural Properties, Durability and Fatigue Life Evaluation (S–N Curves and Constant-Life Diagrams) of Aligned Plant Fibre Civil Composites with Various Configurations and Architecture.
From Everand
Flexural Properties, Durability and Fatigue Life Evaluation (S–N Curves and Constant-Life Diagrams) of Aligned Plant Fibre Civil Composites with Various Configurations and Architecture.
Muhammad Musaddique Ali Rafique
No ratings yet
Autographic Design: The Matter of Data in a Self-Inscribing World
From Everand
Autographic Design: The Matter of Data in a Self-Inscribing World
Dietmar Offenhuber
No ratings yet
X-Ray Fluorescence Spectrometry
From Everand
X-Ray Fluorescence Spectrometry
Ron Jenkins
No ratings yet
The Science of Superhydrophobicity: Enhancing Outdoor Electrical Insulators
From Everand
The Science of Superhydrophobicity: Enhancing Outdoor Electrical Insulators
indirajith K
No ratings yet
The Philosophy of Death: A Simple Guide to Big Ideas
From Everand
The Philosophy of Death: A Simple Guide to Big Ideas
NOVA MARTIAN
No ratings yet
A User's Guide to Ellipsometry
From Everand
A User's Guide to Ellipsometry
Harland G. Tompkins
No ratings yet

4-Evaluation of Dynamic Analysis

Uploaded by

4-Evaluation of Dynamic Analysis

Uploaded by

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Evaluation of Dynamic Image Analysis for Characterizing Granular Soils

Article in Geotechnical Testing Journal · November 2019

Linzhu Li Magued Iskander

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

Evaluation of Dynamic Image

doi:10.1520/GTJ20190137 available online at www.astm.org

Linzhu Li1 and Magued Iskander2

Evaluation of Dynamic Image Analysis

Cx = convexity = Ap/(Ap + As)

Descriptor Formula Description Reference

EQPC de Equivalent circle diameter (ASTM F1877-16)

Pictorial representation of notation

4 4.75 4.75 4 4.75 4 4.75

Quartz #1 20× Quartz #2 20× Quartz #3 20× Quartz #4 20×

GRANULAR SOIL PARTICLE SIZE AND SHAPE DESCRIPTORS

Particle Shape Evaluated

Quartz #3 AR Particle number 2 19 62 144 188 224 208 102 8

Silica #20-30 AR Particle number … … … 2 15 84 164 205 53

Effect of Various Test Parameters on Experimental Results

EFFECT OF PRIMARY AIR PRESSURE

EFFECT OF SAMPLE MOISTURE CONTENT

Actual Water Content

Prepared Moisture Content Silica #20-30 Quartz #3

0.50 % 0.42 % 0.41 %

EFFECT OF SAMPLE WEIGHT ON MEASUREMENT REPEATABILITY

EFFECT OF SPECIMEN SIZE ON TEST DURATION

Repeatability of DIA Test Results

Engineering Performance of DIA

Silica #20-30 Quartz #3

Trial D10, μm D30, μm D50, μm D90, μm D10, μm D30, μm D50, μm D90, μm

1 704.86 799.82 846.00 938.37 760.70 1,009.00 1,163.60 1,550.60

SINGLE GRANULAR PARTICLE SYSTEMS

Standard Standard Standard

BINARY PARTICLE SYSTEMS

Silica #12-20 and #140-270

1:9 5 45 1,420 114 105 87 65 1.75 1.02

Silica #20-30 and Quartz # 12

#20-30 (g) # 12 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

1:9 5 45 773 229 210 178 141 1.63 0.98

Quartz #3 and Quartz # 12

#3 (g) # 12 (g) D90, μm D60, μm D50, μm D30, μm D10, μm Cu Cc

1:9 5 45 766 219 199 168 125 1.75 1.03

Quartz #3 and Silica #140-270

1:9 5 45 1,201 116 106 89 67 1.72 1.02

Silica #12-20 and #140-270

Mix Coarse Fraction #12-20 Fine Fraction #140-270

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 1,000 3,612 1,514 973 147 97 52

Silica #20-30 and Quartz # 12

Mix Coarse Fraction #20-30 Fine Fraction # 12

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 328 1,029 804 684 299 194 103

Quartz #3 and Quartz # 12

Mix Coarse Fraction #3 Fine Fraction # 12

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 280 2,594 1,156 682 299 194 103

Quartz #3 and Silica #140-270

Mix Coarse Fraction #3 Fine Fraction #140-270

Weight Ratio D50, μm D95, μm D50, μm D5, μm D95, μm D50, μm D5, μm

4:6 802 2,594 1,156 682 147 97 52

You might also like