1163
Study on mechanism of two-phase flow in porous media using X-ray CT Image
Analysis
Etude sur le mécanisme de transfert biphasé dans les milieux poreux par l’imagerie aux rayons X
Mukunoki T.
Graduate School of Science and Technology, Kumamoto University, Japan
Mikami K.
Japan Oil Gas and Metal National Corporation, Japan
ABSTRACT: The objective of this study is to understand the mechanism of light non-aqueous phase liquid (LNAPL) migration in
sandy soil. This paper introduces a newly developed test apparatus for measuring flow injection of micro-focused X-ray computed
tomography (MXCT) scanners, and an image analysis technique. The pore scale and its distribution in the specimen are then
evaluated. Only those pore structures trapping LNAPL are extracted, and cluster analysis is applied to each. This image processing
technique allows quantitative evaluation of pore scale and pore structures trapping LNAPL.
RÉSUMÉ: L'objectif de cette étude est de comprendre le mécanisme de migration du LNAPL dans un sol sablonneux. Dans cet
article, la technique récemment mise au point d'analyse de scanner et d'images par appareil de test d'injection de flux micro-focalisé
de tomographie par rayons-
X assisté par ordinateur est décrite. L’objectif est d’évaluer l'échelle de
porosité et sa répartition dans
l'échantillon. Seules les structures poreuses formant les LNAPL sont extraites et une 'analyse de cluster est appliquée à chaque
structure poreuse. Cette technique de traitement par image a permis une évaluation quantitative de l'échelle et de la structure poreuse
du LNAPL.
KEYWORDS: Multipfase flow, ground contamination, Image Analysis, X-ray CT
1 INTRODUCTION
Multiphase flow is an issue in the study of ground
contamination by light non-aqueous phase liquids (LNAPLs),
where quantitative parameters regarding particle shape and pore
structure are often desired. The basic mechanics of LNAPL
migration were developed along with the theory of unsaturated
flow mechanics. In recent studies, van Genuchten and Mualem
models have been used to deduce the relative permeability
coefficient of NAPLs in porous media (Mualem, 1976; van
Genuchten, 1980), allowing visualization of the degree of
saturation distribution for LNAPLs by numeric simulation (e.g.,
EPA, 1997). However, LNAPLs flow though pore structures,
potentially polluting groundwater due to residual LNAPLs in
pores. Characteristics of the pore structure of sand, such as pore
size, shape, spatial distribution, and connectivity, are therefore
important, and should be evaluated in three dimensions to
understand the mechanism of LNAPL capture. Recent
technologies such as X-ray computed tomography (CT)
scanners have made possible nondestructive evaluation of pore
diameter and distribution in the study of fluid behavior in sandy
soil (Wildenschild et al., 2002; Altman, 2005; Wildenschild et
al., 2005; Al-Kharusi & Blunt, 2007; Mukunoki et al., 2010).
New techniques for image processing and analysis of X-ray CT
images are thus now required (Mukunoki et al., 2011).
The objective of this study is to understand the mechanism
of LNAPL migration in sandy soil. This paper introduces a
newly developed test apparatus for injecting fluid for micro-
focused X-ray computed tomography (MXCT) scanners, and an
image analyzing technique proposed by the authors for
evaluating pore scale and distribution. Pore structures trapping
LNAPLs are then visualized by cluster analysis, allowing static
evaluation of pore scale and structure. Specifications of the
micro-focused X-ray CT scanner used in this study are in
Mukunoki et al., (2011).
1 EXPERIMENTAL METHODS
Figure 1 shows the developed fluid injection test apparatus. An
aluminum mold holds the sand sample to be scanned. The inner
diameter of this mold is only 10 mm, making gravity effects
negligible. Fluids are injected using two syringe pumps with a
minimum constant flow velocity of 0.1 cm
3
/h. The developed
experiment consists of injecting fluids into a 14.61 cm
3
sample.
Mass outflow is measured using a precision balance, and the
injection rate is controlled through syringes and a syringe pump
that allow slow rates. Table 1 shows specifications of the fluids
tested. To allow discrimination of LNAPL and water phases
using micro CT, iodine potassium was added to the water. This
KI water has a density of 1.25 g/cm
3
, but its kinematic viscosity
is almost the same as that of pure water. The LNAPL used is
Syringe pump
Pressure
Precision balance
Mold-store soil
(a)
Photograph of test apparatus
(b) Schematic of test apparatus
Precision balance
200mm
Syringe pump
Syringes
Pressure sensor
Figure 1. MXCT fluid injection test system
