3017
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
1
Measurement of NAPL saturation distribution in whole domains by the Simplified
Image Analysis Method
Mesure de la distribution de la saturation de liquide en phase non-aqueuse couvrant tout le spectre
de l’étude par la méthode simplifiée d’analyse d’image
G. Flores
Graduate School of Engineering, Kyoto University
T. Katsumi, T. Inui & A. Takai
Graduate School of Global Environmental Studies, Kyoto University
ABSTRACT: A
Simplified Image Analysis Method
devised to assess the saturation distribution of water and
Non-Aqueous Phase
Liquids
(NAPLs) in granular soils subject to fluctuating groundwater conditions was developed and tested for ten different NAPLs of
different density and viscosity values
(0.73 ≤
ρ
≤ 1.20 g/cm
3
; 1.4 ≤
ν
≤ 1000 mPa∙s)
. The
Simplified Image Analysis Method
, which is
based on an extension of the
Beer-Lambert Law of Transmittance
that predicts the existence of a linear relationship between the
saturation of water (
S
w
), NAPL (
S
o
), and their corresponding average optical densities (
D
i
), was tested by photographing samples of
Toyoura sand mixed with different amounts of water and NAPLs, using two digital cameras with different wavelength band-pass
filters (
λ
= 450 nm and 640 nm), and obtaining the linear equations relating
S
w
,
S
o
and
D
i
for the each NAPL. Once the linear
relationships were confirmed, this method was used to assess the behavior of two different NAPLs subject to fluctuating groundwater
tables, demonstrating that this
non-intrusive
and
non-destructive
method can be used as a reliable tool to provide water and NAPL
saturation distributions in full domains, when studying the effects of porous soil contamination by NAPLs under dynamic conditions.
RÉSUMÉ: Une
méthode simplifiée d’analyse d’image
visant à mesurer la distribution de la saturation d’eau et des
liquides en phase
non-aqueuse
(NAPLs) des sols granuleux soumis aux fluctuations des eaux souterraines, a été développée et testée dans dix différents
NAPLs (0.73≤
ρ
≤1.20 g/cm
3
; 1.4≤
ν
≤1000 mPa∙s). La méthode simplifiée d’analyse d’image, une extension de la
lois de Beer-Lambert
qui établit une relation linéaire entre la saturation d’eau (
S
w
), de NAPL (
S
o
) et leurs densités optiques respectives (
D
i
), a été testée en
photographiant à l’aide de de
ux cameras digitales ayant des filtre passe-
bande de longueur d’onde différente (
λ
= 450 nm et 640 nm)
des échantillons de sable de Toyoura, mélangés avec des quantités différentes d’eau et des NAPLs, et en obtenant les équation
s
linéaires qui lient
S
w
,
S
o
et
D
i
à chacun de NAPL. Après confirmation de la relation linéaire, cette méthode a été utilisée pour évaluer
le comportement de deux différents NAPLs soumis aux fluctuations des nappes phréatiques; et pour démontrer que cette méthode non
intrusive et non
destructive, peut être utilisée de manière fiable pour obtenir les distributions de saturation d’eau et de NAPL dans tout
le spectre lors des explorations des effets de contamination des sols poreux par des NAPLs sous les conditions dynamiques.
KEYWORDS: NAPL, simplified image analysis, saturation, optical density, column test
1 INTRODUCTION
When released in the vadose zone,
Non-Aqueous Phase Liquids
(NAPLs) pose significant contamination risks to the
groundwater (Mercer and Cohen 1990; Capiro, Stafford et al.
2007). Remediation of these releases in an efficient and cost-
effective way should be guided by field data interpreted by
numerical models using the appropriate assumptions
(Kechavarzi, Soga et al. 2000). To verify the accuracy of these
models, laboratory tests should be run and precise saturation
information should be obtained, especially under the dynamic
conditions usually present in nature (Lenhard and Parker 1987;
Fagerlund, Illangasekare et al. 2007; Flores, Katsumi et al.
2011). In this study, we aim to validate the
Beer-Law of
Transmittance
, the basis of the
Simplified Image Analysis
Method
for ten different NAPLs with different density and
viscosity values, and then use this method to assess the behavior
of five different NAPLs subject to fluctuating groundwater
conditions, which may have a significant effect on the behavior
of NAPLs, particularly with regards to their residual saturation.
For this, residual saturation values at the end of drainage and
imbibition stages will be compared for our different NAPLs.
2 SIMPLIFIED IMAGE ANALYSIS METHOD
The
Beer-Lambert Law of Transmittance
states that when a
beam of monochromatic radiation
I
0
strikes a block of absorbing
matter perpendicular to a surface, after passing through a length
b
of the material, its power is decreased to
I
t
as a result of
absorption:
(1)
where
D
i
is the optical density,
ε
a numerical constant,
b
the
length of the path,
c
the number of moles per liter of absorbing
solution,
I
o
is the initial radiant power, and
I
t
the transmitted
power (Skoog et al. 2007). For digital images, the average
optical density
D
i
is defined for the reflected light intensity as:
∑
∑ (
(
))
(2)
where
N
is the number of pixels contained in the area of interest
and, for a given spectral band
i
,
d
ji
is the optical density of the
individual pixels,
I
ji
r
is the intensity of the reflected light given
by the individual pixel values, and
I
ji
0
is the intensity of the light
that would be reflected by an ideal white surface (Kechavarzi et
al. 2000).
It has been shown (Flores et al. 2011) that the
Beer-Lambert
Law of Transmittance
establishes a linear relationship between
optical density and the concentration of a dye:
(3)
Flores G.
Katsumi T. Inui T., Takai A.
