TELKOM
NIKA
, Vol.11, No
.1, Janua
ry 2013, pp. 173
~18
0
ISSN: 2302-4
046
173
Re
cei
v
ed Se
ptem
ber 30, 2012; Revi
se
d No
vem
ber
24, 2012; Accepted Decem
ber 2, 201
2
Solid-Liquid Two-Phase Flow Image Reconstruction
Based on ERT Technique in Microchannel
Yonghong Li
u
1
, Xiantao Wang
1
, Mas
a
hiro Takei
2
1
Hebe
i ke
y
lab
of measurem
e
n
t technol
og
y a
nd inst
rum
ent Yansh
an U
n
ive
r
sit
y
, Qinh
uan
g
dao 0
6
6
004,
Chin
a
2
Colle
ge of Sci
ence a
nd T
e
chnol
og
y Ni
ho
n Univers
i
t
y
J
a
p
an
*corres
pon
di
ng
author, e-mai
l
: liu
yh@
y
s
u
.ed
u
.cn
A
b
st
r
a
ct
Monitori
ng
the
flow
beh
avi
o
ur on
the
mic
r
o-sca
le
is v
e
ry importa
nt i
n
many
in
dus
trial a
n
d
bioc
he
mic
a
l
pr
ocess, the
mul
t
iphas
e co
exist
ence
in
mi
croc
han
nel
prov
id
e
s
many
attractive ch
aracteristi
c
s
compar
ed t
o
a
singl
e-p
has
e fl
ow
. T
he pr
ecis
e flow
rat
e
co
n
t
rol a
n
d
w
e
ll-d
e
fine
d c
han
ne
l
ge
o
m
etries
make
i
t
p
o
ssi
bl
e
fo
r us to
m
a
ke
d
e
t
ai
l
e
d
i
n
ve
sti
g
a
t
io
n
on
mu
l
t
ip
h
a
se
fl
o
w
p
h
e
n
o
m
e
n
a
on
m
i
cro sca
l
e
. Th
i
s
paper
ai
ms at the sol
i
d
-liq
ui
d tw
o phase flow
visua
l
i
z
a
t
i
on i
n
the cr
oss-sectio
ns of a novel
micr
oc
han
nel
base
d
on
the el
ectrical
resistanc
e to
mo
gra
phy (ER
T
) techni
q
ue.
Experi
m
ental r
e
sults rev
eal
that ERT
i
m
a
g
e
reconstruction technique
base on Agilent data
acquis
ition system
c
a
n effectiv
ely detect the
particle
distrib
u
tion i
n
the microc
han
n
e
l.
Key
w
ords
:
two-pha
se flo
w
, m
i
crochan
nel, Agilent instrum
ent, ERT
Copyrig
h
t
©
2013
Univer
sitas Ahmad
Dahlan. All rights res
e
rv
ed.
1. Introduc
tion
In the past d
e
ca
de
s, the n
e
w research
area
of
multipha
se flow
o
n
the micro
-
scale
wa
s
rapidly
devel
oped
du
e to
the a
ppe
ara
n
c
e
of mi
crot
e
c
hn
ology [1,2
] and
availabi
lity of fabri
c
at
ing
compli
cate
d
sen
s
in
g confi
guratio
n o
n
micro sc
al
es usin
g Mi
cro
-
Electro-M
e
chani
cal Syst
ems
(MEMS)
tech
nology. A nu
mber of inve
stigations o
n
microchan
nel
have been p
a
rtially discu
s
sed
in som
e
revie
w
s
acco
rdin
g
to modellin
g
study [3],
flow pattern an
a
l
ysis [4], flow
engin
eeri
ng a
n
d
appli
c
ation
s
:
heat tran
sfer [5], micro
re
actor [6]
a
n
d
DNA a
nalysi
s
[7].
Ho
wev
e
r, the
fluid
s
of
intere
st in
su
ch a
ppli
c
atio
ns a
r
e
ra
rely
singl
e-
p
h
a
s
e
liquid
s
, a cl
e
a
r u
nde
rsta
n
d
ing of the
m
i
cro-
scale effe
cts on multipha
se flo
w
is
still rudime
nt
ary, one of the rea
s
on
s i
s
surely a la
ck of
experim
ental techni
que
s which allo
w
for sen
s
in
g
the
material i
n
ve
ry limited
spa
c
e. Ma
ny opti
c
al
measuri
ng m
e
thod
s such
as mi
cro p
a
rticle
im
ag
e
velocimet
r
y (micro PIV)
and mi
cro la
ser
indu
ced fluo
rescen
ce
(micro-LIF) [8] h
a
ve been u
s
ed attemptin
g
to visuali
z
e the multiph
a
se
distrib
u
tion in
micro
c
ha
nne
l, but these method
s ca
n’
t overcom
e
their inh
e
re
nt optical scatte
ring
probl
em which become
s
e
v
en more
sev
e
re withi
n
the
limited dime
nsio
n.
Process tom
ogra
phy, as
a rapi
d devel
oping vi
suali
z
ing techniq
u
e
,
has b
een
a
pplied to
multipha
se fl
ow me
asure
m
ent within
conve
n
tional
volume u
s
ing
techniq
ue
such a
s
Ele
c
trical
Cap
a
cita
nce
Tomog
r
ap
hy
(ECT
) [9] a
n
d
Electri
c
al
Re
sista
n
ce To
m
ogra
phy
(ERT) [10]. Sin
c
e
it
has th
e adv
antage
s
su
ch as
non
-int
rusive
ne
ss
a
nd excellent
time re
solut
i
on, it has
g
r
eat
potential to b
e
appli
ed to t
he re
se
arch
of multipha
se
flow in mi
cro
c
ha
nnel i
n
th
e future. So
me
attempts hav
e been m
ade
in recent year to visuali
z
e the two-ph
ase flo
w
in m
i
cro
c
h
ann
el u
s
ing
ECT tech
niqu
e[11], but the capa
citan
c
e
-
sen
s
in
g elect
r
ode
s can’t b
e
made very small to give a
sufficie
n
t ca
p
a
citan
c
e
cha
nge; re
si
stan
ce el
ectrode
s can be
m
ade very tiny
and resi
sta
n
ce
informatio
n h
a
s be
en
su
ccessfully retrie
ved [2
], but the mea
s
u
r
e
m
ents b
e
twe
en any ele
c
trode
combi
nation
s
were not fulfi
lled co
ncurre
ntly
inside th
e micro
c
han
n
e
l, so the E
R
T techni
que
has
not been
com
p
letely applie
d in the visual
ization of two
pha
se flow.
In this pa
per,
Agilent
ba
se
d Data
acq
u
i
s
it
ion unit wa
s con
s
tructe
d
for re
si
stivity
data
retrieval from
any need
ed
electrode
pairing combin
ation within
microchan
nel. Th
e co
ncentrati
on
of
ea
ch pha
se
can be co
mputed ba
se
d
on
the kn
o
w
led
ge
of the
ele
c
trical
co
ndu
ctivity of each
pha
se, yieldin
g
the two-p
h
a
s
e con
c
entration tomog
r
am
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 1, Janua
ry 2013 : 173 – 1
8
0
174
2. ERT s
y
stem for microc
hannel
The E
R
T
wo
rkin
g p
r
in
cipl
e con
s
ist
s
of
inje
cting
ele
c
tri
c
al
cu
rre
n
t
betwe
en
a
pair
of
electrode
s a
n
d
mea
s
u
r
ing t
he pote
n
tial d
i
fferences
bet
wee
n
the rem
a
ining el
ect
r
o
de pai
rs. T
h
is
pro
c
ed
ure is
repe
ated fo
r
all the othe
r
electrode
pai
rs
until a full
rotation of the
elect
r
ical fiel
d is
compl
e
ted to form a set of measurement
s. In
our experime
n
t, the
ERT system i
s
com
p
o
s
ed
of
a hard
w
a
r
e p
a
rt, which include
s the se
nso
r
of electrode
s array within micro
c
h
annel, the da
ta
acq
u
isitio
n sy
stem ba
se
d o
n
Agilent me
asu
r
em
ent
in
strum
ent, flow rate
co
ntrol
ling pum
p an
d a
comp
uter
with image reco
nstru
c
tion al
g
o
rithm to re
co
nstru
c
t imag
e
s
.
(a) G
eometry of the chan
nel
(b) Electrod
es
configuration
Figure 1.
The
configu
r
ation
of the micro
c
hann
el
2.1. Microch
a
nnel geome
t
r
y
The micro
c
h
annel [Covalent Material
s Corp
orat
ion,
Japa
n] used
in this rese
arch ha
s
recta
ngul
ar
config
uratio
n
as sho
w
ed
in
Figure 1. On the
quart
z
-gla
ss-made fram
e
w
ork
(
570
μ
m width
), there are totally five
measu
r
ing cro
ss-se
ction
s
, nei
ghbo
ring cro
s
s-sectio
ns a
r
e
sep
a
rate
d fro
m
each othe
r by 4.5mm. In each cr
o
ss-se
ction
16 resi
stan
ce se
nsin
g ele
c
tro
des
are emb
edd
e
d
, which are sep
a
rate
d fro
m
each othe
r by
80
μ
m
. The inlets config
uratio
n used for
microfluids injection is
Y
config
uratio
n
as sho
w
ed
in Fig.1
(b
).
Fluids with
d
i
fferent p
h
ysi
c
al
prop
erty ca
n be inje
cted in
to it and mixed together to
form a two
-
ph
ase flo
w
.
2.2. Data a
c
q
u
isition s
y
stem
The data a
c
quisitio
n
unit’s functio
n
is
to transmi
t the data sampled from
micro
resi
stan
ce el
ectro
d
e
s
into
the comput
er r
api
dly an
d accurately. Convention
a
l ERT syst
em
P2000 E
R
T (Tomog
rap
h
y Comp
any) l
a
ck th
e ability
to
overcome
the interfe
r
e
n
ce
of co
nta
c
t
resi
stan
ce in
the microcha
nnel
ap
plication due to its
poor m
a
ximu
m exciting cu
rre
nt frequ
en
cy
of only 153.6
Hz
. In
our
sy
stem hi
gh-te
ch Agilent me
asu
r
ing i
n
st
ru
ments
wa
s e
m
ployed a
s
d
a
ta
acq
u
isitio
n unit, the Agile
nt 34980A
was used for
cha
nnel swit
chin
g and th
e Agilent E4980
impeda
nce
measurement
meter for
re
al time re
si
stance mea
s
u
r
ement. 349
3
2
A, as a pl
u
g
-in
swit
chin
g mo
dule
of Agile
nt 349
80A pl
atform, was
use
d
for the
co
ntrol
of el
ectro
de
pai
ri
ng
operation a
s
sho
w
e
d
in Fig.3. Every row and
co
lum
n
can b
e
intersecte
d to create the
cro
ss
point. It has e
x
cellent
scan
velocity up to
100
ch/
s
ec
a
nd at the
sam
p
ling fre
que
n
c
y of 1
MH
z
t
he
cro
s
stalk
can
be
ded
uced t
o
-5
5
dB
. T
h
e
s
e
excell
ent
electri
c
al
feat
ure
s
m
a
ke A
g
ilent family v
e
ry
suitabl
e for m
easure
m
ent o
n
the re
sist
an
ce di
stributio
n in the micro
c
ha
nnel.
2.3.
Flo
w
ra
te co
ntrolling pu
mp and computer
for da
ta analy
s
is
IC310
0 micro pump
s
(
K
D
Scientific, USA
) we
re
used to
synch
r
oni
ze th
e syrin
g
e
injectio
n pro
c
ess and
control the flow rat
e
of micr
oflui
d
s in the corresp
ondi
ng inl
e
ts, the flow rate
can b
e
a
c
cu
rately sel
e
cte
d
in a wid
e
rang
e from
0.0
1
⁄
to
2.0
⁄
. The state of the
Agilent instru
ments
were
controlled
an
d sup
e
rvi
s
ed
by a 2.5
GHz
pe
rsonal
compute
r
thro
ugh
GPIB
interface for real time
data pro
c
e
s
sing.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Solid-Liq
uid Two
-
Pha
s
e F
l
ow Im
age Re
con
s
tru
c
tion
Based o
n
ERT … (Yong
ho
ng Liu)
175
3. Image reconstr
uction
for ERT in microcha
nnel
The g
oal of i
m
age
re
con
s
truction
in E
R
T i
s
to com
pute a tom
o
g
r
am
rep
r
e
s
en
ting the
electri
c
al
re
si
stivity distribu
tion of materi
als fl
o
w
ing
wi
thin som
e
col
u
mn fro
m
voltage
s mea
s
u
r
ed
at the periph
e
r
y of the sen
s
or in re
sp
on
se to the inject
ed ele
c
trical current.
To re
con
s
tru
c
t the two phase flow ima
g
e
for ERT in microchan
nel
, forward pro
b
lem and
inverse probl
em have to be solve
d
. K
nowi
ng the resi
stivity distribution
r
an
d given cu
rrent
injectio
n, the probl
em of finding the
ele
c
tri
c
al pote
n
tial
V
inside
or on the bo
un
dary of chan
nel
betwe
en ele
c
trode pai
rs is
calle
d the forward problem
, which
can b
e
denote
d
:
Ex
ci
tem
e
nt C
u
r
r
e
nt&
Excited Voltag
e
Mea
s
u
r
em
e
n
t
Switches&
M
u
ltip
le
x
e
rs
(
A
g
ile
nt 34
98
0)
Figure 2.
Dat
a
acq
u
isitio
n system a
nd e
l
ectro
d
e
s
pai
ring co
nfiguration
Figure 3.
Simplified Sche
m
a
tic of 3493
2
A
switch mod
u
le
()
VF
r
(1)
W
h
er
e
F
i
s
defined
a
s
t
he fo
rwa
r
d
operator co
n
nectin
g
resi
stivity distribut
ion a
n
d
electri
c
al p
o
tential.
The
ch
ang
e i
n
voltage
differen
c
e
s
VV
in re
spo
n
se to
a
pertu
rbatio
n
of re
si
stivity
distrib
u
tion
rr
ca
n be expre
ssed by the Taylor expan
sio
n
:
2
()
(
(
)
)
F
Vr
r
r
(2)
Negl
ectin
g
the highe
r orde
r term
s, Eq.(2
)
can b
e
sim
p
lified to be the linear fo
rm:
VS
r
(3)
Usi
ng finite
element m
e
thod
s (
FEM
),
r
ca
n be
su
bdivided i
n
to
n
di
screte v
a
lue
s
.
Every discret
e
value
corre
s
po
nd
s to on
e pixel
of the
image
re
con
s
tru
c
ted. Sup
pose that the
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 1, Janua
ry 2013 : 173 – 1
8
0
176
are
m
voltag
es, the
n
()
/
SF
r
r
is
an
m×
n
matrix.
S
is calle
d Jaco
bian
ma
tri
x
or
sensitivity
matrix, which is com
puted
based on mul
t
iple solutio
n
s of the
FEM
forward ope
ra
tor.
Finding the
resi
stivity distribution
r
ba
sed o
n
the measured vo
ltages
V
is called
the
inverse p
r
obl
em of ERT.
Based
on th
e previo
us
n
o
tation, the i
n
verse p
r
obl
em is to fin
d
the
inverse of the forwa
r
d op
erator:
1
()
rF
V
(4)
In its no
rmali
z
ed li
nea
r fo
rm, the ERT
inverse p
r
o
b
lem
comp
utes the
inverse of the
Ja
cobi
an mat
r
ix:
1
rS
V
(5)
In gene
ral, th
e pixels of im
age a
r
e
mu
ch mo
re tha
n
t
he voltage
s
measured, th
at mean
s
unkno
wn
re
sistivity values are
more th
an kno
w
n vo
ltage me
asurements. So
dire
ct an
alytical
solution for Eq. (4) or (5) does not
exist.
The ERT inverse problem
is ill-posed. Only
approximatio
ns of
F
-1
or
S
-1
can
be fo
und
by num
e
r
ical
techniq
u
e
s. App
r
oxim
ations of
S
-1
are
comm
only d
e
r
ived u
s
in
g a
lea
s
t-squa
re
s m
e
thod
by
comp
uting a resi
stivity
dist
ribution, whi
c
h
minimizes the
difference be
tween the me
asu
r
ed voltag
es an
d the si
mulated volta
ges:
min
~
2
2
r
S
V
(6)
Both dire
ct and iterative al
gorithm
s can be
formulate
d
using diffe
rent approximations
of
S
-1
, such
like line
a
r b
a
c
k-p
r
oje
c
tion
(LBP), Lan
d
w
eb
er meth
o
d
, Newto
n
-Raph
son m
e
th
o
d
(NRM), an
d
Tikhon
ov regula
r
ization
method.
In our experi
m
ent, we use the widely
-
used
Tikho
nov re
g
u
lari
zation m
e
thod a
s
the tool for imag
e recon
s
tru
c
tio
n
.
The ba
sic id
ea of Tikho
n
o
v regula
r
i
z
at
ion method i
s
to add a b
ound term, which
can
cha
nge th
e ill
-po
s
ed
lea
s
t-squ
a
re
s
equ
ation into
a
well-po
sed
e
quation. Solv
e the mi
nimu
m of
this equ
ation:
)
~
~
min(
2
2
2
2
r
r
S
V
(7)
It is tantamou
nt to solve the equatio
n:
V
S
r
I
S
S
T
T
~
)
(
(8)
The sol
u
tion
of Eq.(8) is u
n
ique if
is positive, and:
V
S
V
S
I
S
S
r
T
T
1
1
ˆ
)
(
~
(9)
w
h
er
e
T
T
S
I
S
S
S
1
1
)
(
ˆ
is the
Tikho
nov
reg
u
lari
zation
ge
nerali
z
e
d
inv
e
rse m
a
trix of
S
, and
is the reg
u
lari
zation pa
ram
e
ter. So an iterative
Tikh
o
nov regul
ari
z
ation method
is defined in the
following form:
)
~
(
)
(
~
~
0
~
1
1
1
0
k
T
T
k
k
r
S
V
S
I
S
S
r
r
r
(10
)
Whe
n
k=1, Eq. (10) i
s
the norm
a
tive Tikhonov re
gula
r
ization meth
o
d
.
The preset stoppin
g
param
eter of
this ite
r
ative method
is:
1
1
~
~
~
k
k
k
r
r
r
(11
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Solid-Liq
uid Two
-
Pha
s
e F
l
ow Im
age Re
con
s
tru
c
tion
Based o
n
ERT … (Yong
ho
ng Liu)
177
Whe
n
is less
than a ce
rtain
val
ue, the iteration sto
p
s.
4. Test res
u
lts
Some experi
m
ents
were condu
cted to p
r
ove t
hat ERT can b
e
su
cce
ssfully ap
pl
ied to
the measure
m
ent of two-p
hase flows in microchan
nel
.
4
.
1
.
Ex
pe
r
i
me
nta
l
pr
oce
d
u
r
e
s
Two
kind
s o
f
microfluid
s were used
in this exp
e
rime
nt. One
of them is polymer
microsphe
re su
spe
n
si
on
(Microge
nics
USA
) with
mean pa
rticl
e
diamete
r
of 2.0
m
. The
origin
al pa
rticle den
sity
of the suspen
sio
n
is 1.0
g/cm
3
and the mi
crosphe
re
ha
s non
con
d
u
c
tive
prop
erty. The
other i
s
0.9
g/cm
3
NaCl
solution
whi
c
h
is u
s
ed fo
r re
feren
c
e m
e
a
s
urem
ent an
d
as
the liquid flow in the liquid-solid tw
o-pha
se flow. In ou
r experi
m
ent,
0.9
g/cm
3
NaCl
sol
u
tion an
d
polymer microph
ere
su
spen
sion
were inje
cted
resp
ectively i
n
to inlet A
and
inlet
C
simultan
eou
sl
y unde
r the
same volu
me
rate of
0.01
ml
/s
. in this
c
a
s
e
, two-phase microflow
was
formed in the
microchan
nel
and inde
pen
dent mea
s
u
r
e
m
ents
were d
one on eve
r
y cro
s
s-se
ction
.
An adja
c
ent current inje
ctio
n and me
asu
r
eme
n
t pattern wa
s u
s
ed.
The fre
quen
cy of the
injecte
d
excit
e
ment curren
t was
set 50
kH
z
at an am
plitude of 10
mA.
Elec
tric
c
u
rrent was
firs
t
injecte
d
into
electrode
pair of 1 and 2. T
he volt
age
s b
e
twee
n the ot
her a
d
ja
cent
electrode
s
were
measured. T
he ex
citemen
t
cu
rre
nt was inje
cted i
n
to
the n
e
xt adj
ace
n
t ele
c
tro
des succe
s
si
vely
and voltage
s betwee
n
the other adj
a
c
ent ele
c
tr
o
d
e
s which are not the current inje
cted
electrode
s
were m
e
a
s
ure
d
. The
sam
e
mea
s
u
r
em
ent strategy wa
s ci
rcula
t
ed until ele
c
tri
c
curre
n
t wa
s injecte
d
into e
v
ery electrod
e. In order
to
redu
ce the inf
l
uen
ce of co
n
t
act re
sista
n
ce,
voltages
on t
he current inj
e
cted
ele
c
tro
des
we
re
n
o
t
measured. Since th
ere
are
16 ele
c
tro
d
e
s
in
each cro
s
s-section, a
nd e
v
ery time of
curre
n
t
inje
cting
13 data of
voltage
can be
obtai
ned. So
16×(16
-
3
)
=20
8
data of voltage can
be mea
s
u
r
e
d
on every
cro
s
s-se
ction
.
The COM
S
OL
Multiphysi
cs
3.5 softwa
r
e
wa
s used to
determi
n
e
the bou
nda
ry voltages tha
t
are gen
erat
ed
arou
nd the
sensor
whe
n
the excitation
curre
n
t si
gn
al
is appli
ed. F
i
gure
4 gives representati
v
e
results fo
r th
e 16
-ele
ctro
d
e
in mi
croch
annel, a
si
ngl
e pa
rticle
wit
h
5
m
diameter l
o
cate
d ne
ar
the top right of the diagra
m
, sugge
stin
g that the invading solid p
has
e can result in the cha
nge of
equip
o
tential
s
dist
ribution.
(a)
No
distur
bance
(
b)
5
m
diameter
solid
parti
c
le
as
disturba
nce
Figure 4. Predicted e
quip
o
t
entials in the
cr
o
s
s-se
ctio
n from finite element mod
e
l
i
ng
4.2. Experimental res
u
lts
Figure 5
sho
w
s th
e voltag
es m
e
a
s
ured
in the
expe
ri
ment, figure
(a), (b
) a
nd
(c) are the
corre
s
p
ondin
g
voltage
s m
easure
d
on
cross-sectio
n
Ⅰ
,
Ⅲ
and
Ⅴ
a
s
example
s
, a
ll of them ha
ve
obviou
s
cha
n
ge
with th
e
controlle
d t
w
o
pha
se
flo
w
i
n
sid
e
the
mi
cro
c
h
ann
el
with the
highe
st
voltage
level
about 3.5
mV
. It prove
s
th
at the
cont
act
re
si
st
an
ce
can b
e
effe
ctively su
ppressed
and th
e me
asured
bo
und
ary voltages
are very
sen
s
it
i
v
e to the
cha
nge
s of
re
sist
ivity distributi
on
in the microchann
el unde
r
the experim
e
n
tal con
d
ition
.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 1, Janua
ry 2013 : 173 – 1
8
0
178
(a)
Voltag
es measured
in cro
s
s
section
Ⅰ
(b) Voltag
es
measured in
cro
s
s section
Ⅳ
(c) Voltage
s measured in
cro
s
s section
Ⅴ
Figure 5. Voltage
s mea
s
u
r
ed in som
e
sample cro
s
s se
ction
s
(a)
Re
con
s
tru
c
te
d image on
cross-se
ction
Ⅰ
(b)
Re
con
s
tru
c
te
d image on
cross-se
ction
Ⅲ
(c) Re
con
s
tructed ima
ge
on cross-sect
ion
Ⅴ
Figure 6. 2-D and 3-D Re
si
stivity distribu
tion on som
e
sampl
e
cro
s
s-se
ction
s
0
52
104
15
6
20
8
0
2
4
vi
1
(
m
V
)
t
i
m
e
s
o
f
m
e
as
ur
e
m
en
t
0
52
104
156
208
0
2
4
vi
3
(
m
V
)
t
i
m
e
s
of
m
eas
ur
em
ent
0
52
104
15
6
20
8
0
1
2
vi
5
(
m
V
)
t
i
m
e
s
of
m
eas
ur
em
en
t
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Solid-Liq
uid Two
-
Pha
s
e F
l
ow Im
age Re
con
s
tru
c
tion
Based o
n
ERT … (Yong
ho
ng Liu)
179
4.3 Image recons
truc
tion
The
cha
nge
s of th
e resi
stivity distribution
(
r
) in
every cross-se
ction were
recon
s
tru
c
ted
ba
sed
on
the me
asu
r
ed
bou
nda
ry voltages
usin
g the it
erated
Ti
kho
nov
regul
ari
z
ation
algo
rithm
m
entione
d a
b
o
v
e. Cho
o
se
=0.01, an
d
stoppin
g
p
a
ra
meter
not
greate
r
th
an
0.001 to
fa
cilitate the
co
nverge
nce
of
n
on-lin
ea
r inve
rse
p
r
obl
em.
In this case t
he
2-D a
nd 3
-
D
reco
nstructe
d
image
s of the
r
in some cro
s
s-sectio
ns
were
sho
w
n in
Figure 6 as
example
s
.
Figure 6 reve
als the
cha
n
g
e
s of re
si
stivity distribution in all of
t
he f
i
v
e
cro
s
s-
se
c
t
ions in
the microcha
nnel. We
can
see that alon
g
the flow direction fro
m
cross se
ction
Ⅰ
t
o
cro
s
s
-
se
ct
io
n
Ⅴ
,
Polymer mi
crosp
here susp
ensi
on an
d
NaC
l
solution resolve
into e
a
ch other gra
dually
exce
pt
in
cro
s
s-
se
ct
ion
Ⅰ
,
NaCl
solu
tion just mee
t
each othe
r. The te
st re
sult also
rev
eals that un
der
certai
n flow rate, polymer
micro parti
cle
s
hav
e alread
y disperse
d e
v
enly near cross-se
ction
Ⅴ
.
The recon
s
tructed im
age
s of
r
on every
cro
s
s-se
ctio
n sh
ow
that
the polyme
r
micro
particl
es see
m
to h
a
ve t
he ten
den
cy
to di
spe
r
se
evenly
with
in the li
quid
mediu
m
. T
h
is
phen
omen
on
agre
e
s
with the theory whi
c
h cl
aims t
hat
when two different ph
ases
are inje
cted a
s
adja
c
ent
stre
ams int
o
on
e ch
ann
el, one p
h
a
s
e
will often e
n
c
ap
sul
a
tes t
he othe
r p
h
a
se.
Although th
e
test results
is not ve
ry satisfacto
ry, it agree b
a
si
cally with the
real
experi
m
ent
con
d
ition which sugg
est
s
that the Agilen
t
based me
asuring
system
can really wo
rk.
5. Conclusio
n
We use syrin
ge
inje
ction
method
to
ge
nerate
a
stab
le liquid
-
solid
two-pha
se
flow i
n
a
novel expe
ri
mental mi
cro
c
ha
nnel fo
r i
m
age
re
con
s
truction
of the two-pha
se
flow on th
e micro-
scale. T
he vo
ltage
cha
nge
on the
mi
cro
sen
s
o
r
can
b
e
effectively o
b
tained
an
d fl
ow
phe
nome
na
of the mi
crofluids can b
e
detecte
d u
s
in
g the
High
T
e
ch
Agilent
measurement
strate
gy tog
e
ther
with ce
rtain i
m
age recon
s
truction al
go
rithm. The image of solid
-liquid two ph
ase flo
w
wa
s
su
ccessfully
re
con
s
truct
ed u
s
ing
A
g
ilent dat
a
acq
u
isitio
n system, the
pre
s
en
ce
of
the
non
con
d
u
c
tive micro
s
ph
ere in the
microfluid
s
c
an
alter the
re
si
stivity distrib
u
tion in
side
the
microchan
nel
.
Since
all
of the exp
e
rim
e
ntal results
a
r
e ba
sed
on
non-co
ndu
cti
v
e
micro
s
ph
e
r
e with
same
p
a
rticl
e
si
ze
and
sa
me flo
w
rate,
more
detail
e
d
re
se
arch
sh
o
u
ld b
e
d
one
u
s
ing
pa
rticle
s
o
f
different p
h
ysical p
r
op
ert
y
unde
r othe
r different
te
st condition,
mean
while, t
he real time
3-
dimen
s
ion
dy
namic imag
e
re
con
s
tru
c
tio
n
ca
n n
o
t be
reali
z
e
d
up
to no
w d
ue to
the ha
rd
wa
re
set
up r
e
st
rict
i
on.
Ackn
o
w
l
e
dg
ements
The
present work was suppor
ted by
Japanese research
grant of JSPS (No.21360088)
and the S
c
ie
nce
and
Te
chnolo
g
y Dev
e
lopme
n
t Pro
g
ram
of He
b
e
i Chi
na (No
.
11213
586
), the
authors woul
d like to thank the Coval
ent Materi
al
s Corp
oration
for helping
of micro
c
h
a
n
nel
fabrication.
Referen
ces
[1] Kikutani
Y,
His
a
moto
H, T
o
keshi M,
et al.
Micro wet analys
is system us
in
g multi-phas
e lam
i
nar flows
in three-
di
me
n
s
ion
a
l
microc
h
ann
el n
e
tw
ork
.
Lab C
h
ip. 2
004
; 4: 328-33
2.
[2]
Choi JE, T
a
kei M, Doh
DH.
F
abricati
o
n
of Microc
ha
n
nel w
i
th
60
Electrod
es a
n
d
Res
i
stanc
e
Measur
e
m
ent
.
F
l
ow
meas
ure
m
e
n
t and i
n
strumentati
o
n
. 20
10; 61: 27
33-2
738.
[3]
Reddy
V, Zahn JD.
Int
e
rfacial
stab
ili
z
a
t
i
on
of or
gan
i
c
–aq
ue
ous tw
o-ph
ase
micr
oflow
s
for a
mi
niat
uri
z
e
d
D
N
A
extraction mo
du
le
, Col
l
oi
d Interf. Sci. 2005; 286: 1
58–
1
65.
[4]
Shui
L
L
, Eijk
el
Jan
CT
, Albert
Vand
en
berg.
Mu
l
t
i
p
h
a
s
e fl
ow i
n
m
i
cro
-
and
na
no
ch
an
ne
ls, Se
n
s
o
r
s an
d
Actuators
. 200
7; 121: 26
3–2
7
6
.
[5]
Kaba
n’kov ON,
Sevast’
y
an
ov AP.
Tw
o-phas
e flow
s: a review
, Heat T
r
ansfer
. 2000; 3
1
: 103–
12
2.
[6]
Doku
GN, Ver
boom
W
,
Rei
n
hou
dt DN.
On-
m
icr
o
chi
p
multi
phas
e c
h
e
m
ist
r
y—a revi
ew
of
microre
a
ctor
desi
gn pri
n
ci
pl
es and re
ag
ent
contacting
mo
des
, T
e
trahedron. 200
5; 61: 2
733
–2
742.
[7]
T
egenfeldt JO, Prinz C, Cao
H et al.
Micro-
and n
a
n
o
flui
dic
s
for DNA ana
l
ysis
, Anal. Bio
ana
l. Chem.
200
4; 378: 16
7
8–1
69
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NIKA
Vol. 11, No
. 1, Janua
ry 2013 : 173 – 1
8
0
180
[8]
Ichiy
a
nagi M, Sasaki S, Sato Y. Micro-PIV/
LIF
measure
m
ents on el
ec
troki
netic
all
y
-
d
r
i
ven flo
w
in
surface mod
i
fie
d
microch
ann
el
s,
Journal of Mi
cromech
anics
and Micr
oen
gi
neer
ing
. 2
009;
19: 1-9.
[9]
W
ang Cui
p
i
ng,
Lv Z
i
an, Li Di
ngka
i
.
Experi
m
ental study o
n
gas–so
lids flo
w
s in a circula
t
ing flui
dise
d
bed us
in
g elect
r
ical ca
pacita
n
c
e tomogra
phy
, Po
w
d
er T
e
chnol
og
y. 20
08; 185: 14
4-1
51.
[10]
Ricard
F
,
Bre
c
htelsb
au
er C,
Xu
XY. Mo
ni
to
ring
of Multi
phas
e Ph
arma
ceutica
l
Proc
e
sses Usi
n
g
Electrical R
e
si
stance T
o
mograph
y,
Ch
e
m
ica
l
Engi
neer
in
g Rese
arch an
d
Desig
n
. 20
05; 83: 794-
80
5.
[11]
Sung
Quek,
Stepha
n Mo
hr
, Nick God
dar
d, Peter
F
i
e
l
d
en a
n
d
T
r
evor York. Mini
at
ure El
ectrical
T
o
mograph
y fo
r MicroF
lui
d
ic
S
y
stems,
6th
W
o
rld Co
ngres
s on Industri
a
l
Process T
o
mo
grap
hy
. 201
0;
132
0-13
26
Evaluation Warning : The document was created with Spire.PDF for Python.