A B S T R A C T
A M A
ND A L
DY E
: Lat t ic e Bol tzm
a n nS
im
ul
at ion of
N on-D a rcy F low in
S
ph
e reP
ack ings(
Und
e r th
ed
irect ion ofD
rC
a s s TM
ill
er)
Lat t ic e-B
o
l
tzm
an nm
od
el
ing was us ed to sim
ula te singl
e-p hase, non -D a r c
y f
l
ow t hr ough fl
ow
d
om
ains c onstructed ut il
izing a n al go ri thm
t hat sup po rts th
e ge ne rat ion of
a system
of
sph
e resw i t h log-nor
m
all
yd
istri buted rad
i i a nd
ad
efi
ned
por o si t yR
epre sentat ive el
em
e nta ry v ol
um
e swere
d
eterm
in ed
using t he pe rm
eab
il
i t y and
in ert ial
c o effi
cie nts of
a n on-d
im
ensio n al
m
om
e ntum
equ at io n
f
or a ra nge of
sph
ere-size
d
istrib
ut io nsP
oro u sm
ed
ia were gen erated f
or po r o si ti
e sr ang ing fr o
m
0 3 to 0 6fo r
the d iffe r
e nt sp he r e-siz e d istri but ions T hepa r a
m
ete r s ob
tained
fr o
m
th
e n o n-Da rcy
f lo
wm
od
el d
id
notfi
t the
e xist ing pe rm
e ab
i l i t y and
in ert ial c or relat ionspr opo s ed b y
C
a rm
e n-K
o ze ny,
R
um
pf
-G
up te,Ergu n a nd ot he r s Bas e
d
on t he expe rim
ental d
atac o
ll
ected fr om
th
el
at t ic e-B
ol
tzm
an n sim
ul
at io ns, em
p irical
c o r r ela
t ions a re sug gested
th
at r el
ateA C K N O
W
L ED G
M
EN T S
F ir st I wa nt to t ha nk
m
y ad
vis o rD
rM
i l ler and
DrG
r ayfo r
tak
ing a chanc e onm
eI
d
eep l y ap pr e ciate th
eir pat ie n c e and co ntri but io ns of t im
e, i de a s, andf
u nd ing tom
akem
yM
a ster s e xperie n c eb
oth
prod
uct ive and
st im
ul
at ingI
wo uld
al
sol
ik
e to th
a nk D
rV
izu etef
orta
k
ing t he t im
e to be onm
y t hesis c om m
i t te eI a
m fo r e v e r
ind
eb
ted
toJ
am
esM
cC l
ur ef
o r al l
th
elo n
g-suffe r
ingh
eh
ad
to go th
r ough
toh
el pm
e u nd
er stand
th
e pr ogr am m
ing sid
e of m
y rese arch He a ut hored m
uch
of
th
e cod
ing th
atwent into
m
ak
ing th
is work
pos sib
le and
to ok on t he u noffi
cial
r ole
of
prob
lem
f ixer along t heway
I
am
also gratef
ul toS
ar ah Sh
el tonf
o r al l
th
e em
ot io nal
sup port,
h
el p, a nd
c om
r ad
e rie sheha s pr o vi
d
ed
t hr oug ho ut t h is s om
et im
e s d iffi
c ul t pr o c e s sI
jastl
y,I
wo uld l
ik
e to th
a nk m
yf
am i
l yf
or all
th
eirl
ove a nd
e n c o uragem
e ntF
orm
y sib U
n gs,S
am
ant ha , Eric a,an
d
Kevin,m
y papa,B
il l
, wh
oha s
b
e enm
yb
i g ge stf
an since t heb
eg in ning, al
ways insp iringm
e tof
ol lowm
yd
ream
s, and m
o st of
al lm
y pa r ents. B i l l and
Pam
, wh
oh
aveb
e en a c onstant s our c e of
sup port - em
ot io nal,
m
o ral, and
of
c our se fi
n a ncial
- th
r oug
h
outm
ystu
d
ent ca r e e r a nd
th
is th
esis wo ul d
certainl
y n oth
a v e existed w
i th
o ut th
em
I
t is t ha nk
s to t het wo o
f
th
em f
or ex posingm
e to th
e wo rld
a nd
i gni t ingm
y ongoing se archf
or kn owle
d
ge -i t is toT
A
B L E O F C O
N
T E
N
T S
L
I S T
O F T A
B LE S
vL I S T
O F F I G U R E S
vi
Intro
d
uct ion 1Ba c
k
gro u nd
1M
eth
od
s 4G
ene r at ion ofM
ed
ia 4M
ul t i-Relax at io n T im
e Lat t ice Bol tzm
an nS
chem
e 5R
e sul
ts and
D is cus sion 6D
evelopm
ent of R
epr e se ntat iveE
lem
enta ryV
olum
e s7
D
eriv ed C
or r elat io ns8
Pe r
m
eab
il
i t yC
o r rel
at ion8
Ine rt ia
l C
o r r el
at ion 1 0C
oncl
usio n 1 1L I
S T
O
FT A B
L ES
Ta
bl
e1
G
ene r atedM
ed
iaL I
S T O
F F IG U R
ES
F
i gure1
C
o ef fi
cie nt v al
ue s relat ive to th
e a s s o ciated
R E V v alu
e s ob
tain ed fo
r the pe rm
e¬a
b
i l i t y and
ine rt ial pa r am
ete r sfo
r a r ange of
sph
e r e pa ck ings 72
C
om
pa ris on bet we en th
e simul
ated da
ta a nd
t he ex ist ing em
p iric alrelat io ns 83
Th
e pr opo s ed
pe rm
e ab
ili t y c o r r ela
t ion pr ed
icts t he sim
ulat iond
ata 94
C
om
pa riso nb
et ween th
e simulated d
ata and
t he e x ist ing em
p iric al
r elat ions 1 0I
n
t
ro
d
u
ct
i
o n
S
ing le p ha s efl
ow in por o u sm
ed
ia p lays a fu nd
am
en tal r ole inm
any n atural and industrialpr o c e s s e s
W
h
il
e th
e prim
a ry ap pro a ch
esf
ord
es c rib
ing th
es e system
s are ro o t ed
in em
p iricism
,
r e c ent wo rk has fo cuse
d
on de riving tr ad itional m
odels fo r sing le p ha s efl
ow fr om fi
r st princi p le s and
l ink ingm
a cr o s c op ic ph
e n om
e n a w i th m
ic ro s c alefl
owb
eh
a vi
o r(
H
ass a nizad
eh
a nd G
r ay, 1 9 8 7, 1 9 8 8 ;M
a and R
uth
, 1 9 9 4)
T he cur re nt stat e of know led
ge categorizes sing le p ha s e flow into t wo prim
a ry reg im
e s: a we ak ine rtia r eg im
e c o r r e spond ing to sm
al l fl
ow v elo ci t ie s in w h ichv is c o u s
f
o r c e s c om p l
etel y d
om
in ate in ert ial f
o r c e s a nd
a str o ng in e rt ia r eg
im
e corr e spo nd
ingto
m
od
eratefl
ow vel
o ci t ies in wh
ich
in ert ial f
or ces c a n n otb
e n egl
e cted
(
M
a a nd R
uth
, 19 9
4 ;W
a ng, T hauvin and M
oh
a nt y, 19 99
; Four a r et al
, 2 0 0 4 ; Panfi l
o v and F
o ur a r, 2 0 0 6)
T
he se t wor eg i
m
e s a r e l inked
to t wo of
th
em
o stc om m
onl y ap p l iedm
odels,D
a r cy 's
l
aw
a nd t he Fo r chh
eim
e requat ion
S
tate of
th
e art n um
eri
c al m
eth
od
s pr ovid
e th
e op po rtu ni t y to stud
yfl
ow in poro usm
ed
iad
ir e ct l yf
r om
t hem
ic ros c ale T h ispr o vid
e s th
e op po rtu ni t y to a s s e s sdetail
s of
th
em
ic r o s c alefl
owstru cture in a
d d
i t io n tom
acr o s c op icf
o rm
s Th
el
at t ic e Bol tzm
annm
eth
od h
a sgained
popul
a ri t yf
or sim
ulat io n of
po r o u sm
ed
ia a nd
oth
e r c om
p lex system
sd
u e to c om
putat io n al
ad
v a ntagesov er
m
ore tr ad
i t io n al fl
uid d
ynam
ic s ap pr o a ch
es, pa rt ic ula rl
y inh
a nd l
ing com
pl
e xb
o u nd
a ryc o n
d
i t io ns a nd
parall
el
iz at io nM
ul
t iR
el
a x at io n T im
e(
M R T
)
lat t ic eB
ol tzm
an n schem
es hav eb
ee nd
em
onstr ated
toh
avem
a ny ad
v antage s o v er th
em
or e w id
el y u sed
singl
e r el
axat ion t im
e(
B G K
)
l
at t ic e Bol
tzm
a n nm
od
el
, incl
ud
ing im
pr o ved
n um
erical
stabi li
t y a nd
sim
ul
at io n of h
i gh
R
eynold
s n um b
e rfl
ows(
P
a n, Lu o andM
i l le r, 2 0 0 6)
T
h
e o ve r al l
go al
of
th
is wo rk
is to ad
v anc e t he im d
e r sta nd
ing of
singl
efl
ui d p ha sefl
owb
ype r
f
orm
ingm
icr o s c al
efl
ow sim
ulat io ns using th
el
at t ice Bol tzm
a n nm
et hod
Is otrop ic poro usm
ed ium
syst em
s w i l lb
e co nsid
ered f
orR
eyn old
s n um b
e r s < 1 5 0 Th
e spe cif i
c ob
j
ect iv es of
t h iswo r
k
are;1 to investi gate t
h
em
ost ap pr opriatef
o rm
of
ap pro xim
atem
om
e ntum
equat ions;and2
tod
e v el
op rel
ati
o n s to estim
atem
om
e ntum
equ at io n param
ete r sf
r om
po r o u sm
ed
ium
c
h
a r a cte rist icsB
a
ck
g
ro u n
d
D
a r cy'sl
awf
or o n e-d
im
ensionalfl
ow inpo r o us
m
ed
ium
c a nb
e e xp
r e s s ed
in th
ef
orm
(
G
raya n
d
M
il
ler, 20
0 6)
:
w he r e p™ is t he
m
a c r o s c alefl
ui d pr e s sur e, p™
is t he
m
a c r o s c ale flui d density, g™
is the ex t ernal fo r c e pe r u ni t
m
a s s a ct ing on th
efl
ui d p ha s e in t he x d ir e ct ion, e is t he po r o si t y, j ;™ is t he
b
a ryc e ntricm
a cro s c ale vel
oci t y of
th
efl
uid rel
at ive to th
e s ol id
in t he xdi
rect io n , and R
is t hem
om
entum
resista nc e c oef fi
cie nt Th
efo
rm
of R
is of
princi
ple
concer nf
o r th
e stud
y of fl
ow inpo r o u s
m
ed ia and isk
nown tod
epend
on pr ope rt ies of the
por o u sm
ed
ium
a ndfl
uid a s wel l a stli e
R
eyn ol ds n um
be r :j
l" ' ^ 'w
h
e re /!
"'
is t
h
ed
yn am
ic v is cosi t y a nd d
is a ch
ar acte rist ic po rele n
g th
s cale
Forfl
ows in wh
ich
ine rt ia
l f
or c e sm
ay n ot be negl
e cted
, t hem
om
entum
r e sista nc e co effi
cient c anb
e ap pr o ximated
b y
(
M
cC lu
r e,G
r ay and M
i l le r, 2 0 1 0)
:R
=^
^
a{
dn)
+b
{m ) \
R
e\
Y
(
3)
T
h
e c oeffi
cie nt a is related to th
e in trinsic perm
e ab
ili t y of
th
e po r ousm
ed
ium
a nd b
rel
ates to the streng t h of
ine rt ial effe cts Th
es e c o effi
cie nts pr o vide am
a c r o s c op icm
e a sur e of
th
e infl
uenc ethat porou s
m
edium
m
orp hol
og y and
topol
og y e xert o nm
ic r o s cop icfl
owb
eh
av ior, a nd
are t he r ef
or e pr e sum
ed to depend o n an u nknow n s et ofd
imensionle s sm
o rp holog ic alm
e asur e s9H
I
d
e nt ifi
c at ion of
an ap propriate setofm
easur e spe rm
i tsth
e co nstruct ion of
pred
ict ive co nst i tut ivel
aws, an ob
j
ect iv e w h ich ha sb
e en pur sued
e xte nsiv el yb
y oth
e r aut ho r s(
Tek
, 195
7
;G
e ertsm
a,1 9
74
)
I
n o rde
r to sim
pli
f
y s ub
s equ e nt a n al
ysis, ad
ime n sio nle s s
fo
rm f
o r th
em
om
e ntum
equatio n
ca n be c onstr u cted
by
substit ut ingE
q 3 intoE
q 1 and m
ul
t i p l y ing b y p" ' (
f
/ {
p
^'
)
to ob
tain:F
o =[
a(
Dn
)
+b
(O
T ) \
R
e\
\
R
e,(
4
)
w
h
e r e th
ed
im
ensionl
e s sfo
r cing te rm
isd
efi
n ed a s:F
^ =7
^
{
-^
+ P'"
9
^
]
(
5)
T
h
em
o st si gnificant a spe cts of
poro usm
edi
um m
orpho
lo
g y w i th
respect to pr edi
ct io n of
t
h
e c oeffi
cients a and b
, a re th
e porosi t y e and
t he spe cifi
c s urface a re a e *Su
rf
a c e-to-v ol
um
erat io p
la
ys a key role
inm
ic ros cop icfl
ow proce sses, a nd
i ts ef fect ca nb
e incorporated
into t hed
imensionl
e s sfl
ow equat ions b y rel
at ing i ts val
ue to th
e leng t h s c alefo
r t he system
:T he length defi n ed
b
y E q 6 is the d iam
ete r of t he sp he r e w ith the equiv alen t s u rfa c e-t ov olu
m
er at io, t y p ic al l y known a s t he
Sa
ute rd
iam
ete rW
i t h t he ef fe ct sd
u e to t he spe ci f ic surf
a c e a r ea a c c ou n ted fo r b y th
el
e ng t h s c aled
ef inition,e xist ing c o r r e
l
at ions t y pi
c al l
y pr ed ict th
e co effi
cie nt valu
e s a a nd b
asf
u n cti
o ns of
th
e po r o si t yonl y T
h
em
o st wel l
-kn o
w n of
t hese e xpressions isE
rgu n 's equat io n, w
h
ich
spe ci fi
e sf
unct io nalfor
m
sf
o r th
ed
im
ensio nle s s perm
e ab i l i t y:'
-
~
.
=
^
^
^
^
^(
^)
an
d
in ert ial
c o effi
cie n
t:b =
B
i
l
: :
i
l
(
8)
W
h
e n t hese expres sio ns are com
pared
to th
os e as sociated
w i th
t he volu
m
etricfl
ow r ate, u s e of
t
h
eR
eyn old
s n um b
er a sd
efi
ned b
y E k j 2 a c c ou ntsfo
r ad
iff
e r ence of
1/
e in t he pe rm
eab
i l i t ye xpr e s sion an
d
1/
e^in
th
e e xpres sio nfo
r th
e ine rt ial
co effi
cient Ergun sug ge sted
th
atA
=150
a n
d
B = 1 7 5(
Ergu n, 19
5 2)
,b
ut se v erald
i ff
e r ent v alue sf
o r the
c o ef
fi cients ha v e be en pr opos ed
W
h
enA
— 18
0,E
q7
is th
e w id
el
y us ed C
a rm
e n-
K
o ze ny relat ion sh i p(
B
e a r, 19
72
)
M
a cD onal d
et a
l d
eterm
in ed
a r a ngef
o r the B c oeffi
cie nt, 1 8 <B
<4 0
,th
at ser v ed
tom
atch
experim
e ntal
d
ataf
rom
sixd
if fe
re nt poro u sm
ed
ium m
od
el
s(
M
acd
o n al d P
, 19 79
)
O
th
e r co r rel
at io nf
orm
sh
aveb
ee n propos ed
ad d i
t ionall
y to th
esem
od
ifi
ed
Ergu nf
o rm
sFo r t
h
e c a se of D
a r cyfl
ow ,R
um
pf
and G
up te pr ed
ict th
e pe rm
e abi l i
t y using th
e r el
at ionsh
i p(
R
um
pf
andG
up te, 19
7 1)
:—
= -i
- e^ '^(
9)
D
^ 5 6^
^>Pa n et al
f
ou nd
th
at th
eC
a rm
en- Ko z eny r el
at ionsh
i p u nd
e r e st im
at e s t he pe rm
eab i l i t y, a nd
o
b
serv ed d
ev iatio n s
fr o
m
th
eR
um
pf
-G
up te re
l
at io n whe n
system
s o utsi de
th
e ra nge of
e xper¬i
m
e ntal
sup po rtfo r
th
is e x pres sion we re co nside
red
Th
ey pr oposed
a n al tern a tive cor rela
t io nfo r
m
in wh
ich a n ad d
i t io n al d
im
e n sio nle s s v a riabl
e, th
e r ela
t iv e sta nd
a rd d
e viat io n a o ,wa s als o
inclu
d
ed
(
Pan,H
il
pert and M
i l le r, 2 0 0 1)
:
-^
=/3
i e ^=
(
l +A
3 a^
^)
(
1 0)
M
o r e gene r al l
y, th
e coef fi
cients in E q 4m
ayd
epe nd
upon any ind
epe ndent,d
im
ensio nl
e s sm
easur e of
po r ou sm
ed iu
m m
orph
olog yF
o r th
e c a s e of fl
ow inh
eteroge n e o us, is otr op icm
ed
iaco
m
posed
of sp here pa ck
ings,we pr e sum
e th
at t fi efl
owm
ayb
e de scrib
ed
ad
equatel y b y a s s um
ingd
epe nd
e nc ie s of
th
efo r
m
:b
(m )
^ bj{
t)
b2
{a D
)
,(
1 2)
w he r e :
a o -
^
-^
L
_ ,(
13
)
in t
h
e c a se t hat th
e r ad
i i a r el
og-n or
m
al l
yd
istrib
uted
w i t hm
e an fi and
v a rianc e a'
^
M
et
h
o
d
sG
ene r at
i
o n of
M
ed i
aSu
r r ogate po ro usm
ed
ia we r e c onstructed
ut i l izing a c ol le ctiv e r e a r r angem
ent al
go ri t hm
tha
t gene rate s sph
e re pack
ings w i t hl
og-no rm
al l y
d
istrib
uted
r ad i i and
ad
efi
n ed
po r o si t y T he al
go¬r
i
th m
isb
a s ed
o n o n e u s ed b
yW
ill ia
m
s a nd P h
il
i pse to generate pa ck ings of
sph
e r e o cyli
nd
e r s(W
il l
iam
s a ndP h
il
i pse, 2 0 0 3)
and f
ol l
ows th
e s em
aj
o r steps ;1
A
system
c ontaininga setn um b
er of sp he r e s of
th
e spe cifi
ed v a riance is c onstructed
w i th
in th
e all
o c ated
dom
ain2
O
v e rlaps bet we e n th
e sp her e s ar e el
im
inated
3
T he siz e of
th
e r ad
i i is in cr e a sed b
y a c onsta ntfa
cto rS
teps2
a nd 3
a re r epeated
u nt il
th
e system
of
packe
d
sph
er e s rea ch
e s th
ed
esir ed
po rosi t yB
y ini t ial
izing and
r e s cal
ing th
e rad
i i in a n ap propriate way,sphe
re pa ck
ingsm
aybe
ge n erated
wi
th
lo
g-n orm
all
yd
istrib
uted
rad
i i:l
og(
r«
)
~ No rm
al
(
M,( T2
)
(
14
)
T he va rianc e c r^ is prov i
d
ed
as a n input pa r am
eter wh
ile
th
em
e a n n isd
eterm
ined b
y t hefi
n al
pa ck ing po r o sity
To a c c ele rate c o nvergenc e
,the syste
m
of
sp he r e s isd
iv id
ed
into a s et of cel ls ea ch c om
po s ed of
equall
y-siz ed
s ub d
om
ainsE
a ch of
th
e c el l
s c ontain s al
ist of
th
e sph
e r e s w i t h ce ntr oid
sl
o cated
w i t
h
in th
eb
ou nd
aries of
th
e c el lF
or any g ive n sph
e r e overl
aps are o nl y con sid
e r ed f
r om
th
e nei ghb
o ring cell
, th
eref
ore t he al go ri t hm
runt im
e is acc el
e r ated b
y inc r easing t he n um b
er of
c e
l
ls be cau s e th
el
e ng th
of
th
e se ar ch
path
wh
e n c om
put ing overla
ps isd
e c r e ased
To pr e ve ntov e r si g
h
ts in th
e o v erl
ap com
putat io n in th
e e v e nt the
m
axim
um
sphe
re rad
ius exce eds o n e
h
al f
t he c el l w i
d
t h, th
em
a x im
um
rad
ius is checke
d
af
ter the
sim
ulat io n is com
ple
te a nd
aw
arningM
ul
t
i
-R
el
ax
at
i
o nT i
m
eL
at t
i
c eB
ol
t
zm
a n nS
ch
em
eIn t h iswo rk,we u tiliz e a t
h
r e e -d
im
ensiona
l
,nineteen vel
o ci t y v e ctor(
D
3Q
1 9)
m
ult
i
-relaxat io n
t ime
(
M RT
)
f
o rm
ulat ion of
the lattic eB
oltzm
an nm
et hod to ob
tain ste ad
y state vel
o ci t y f iel dsf
o r a s equen c e ofR
eyn ol d
s num b
e r s In t h is ap pr o a ch, a s olut io nf
o r th
e po r e-s c a
l
e v elo ci t yfi
el d
u
(
x j)
is ob
tained
at e v enl y spa c el
at t ic e si te s x ; b y c o n sid
e ring t he e v olut ion of
a set of di
s c r eted
istribut ions/
,)(
x i)
, w her e th
e ind
ex g is a s s o ciated
w i t h a pa rticula rd
is c r ete v el
oci t y:{
0
,0
,0
}
^
f
or g= 0{
± l ,0,0}
^
,
{
0,
± l ,O
p
'
,
{
0
,0,± l}
'
^
f
o r 9 = 1,
2,
,6
(
1 5)
{
±1,
± 1,0
}
^,
{
± 1,0
,± 1}
'
^
,
{
0,± 1,± 1}
^
f
or g =
7
,8
, ,18
T he density an
d m
om
en tum
ca nb
e expr e s s ed a sl
ine a r c om
b inat ions of
t hed
is c r eted
istribut ions :18
8= 0
18
i
E v
<;(
1 7)
^
« = 0
W
i th
in th
eM
R
Tf
r am
ewo rk
, th
erela x at io n o
f
th
ed
is c r eted
istributi
o n s tow
ard
th
eir equilib¬riu
m
val
u es ism
od
el
ed b
y co nsid
ering a s et of
equili
b
rium m
om
e nts/
„, ob
tained b
y al
in e artra nsfo r
m
at ion of t hed
istri but ions :1 8
0
T
h
e values tr ansf
o rm
at ionm
atrixM
, , ,,, a r e ch
o s enb
ased
on aG
r am
-
S
ch m
idt
ort hogo n ali
z at io nc onstr u cte
d
using pol yn om
ials of t hed
is c r ete v elo ci t ies^
„S
olut ion for t hed
is c r eted
istri but ionsis provi
d
ed b
y a s sum
ing
th
at e a ch m
om
ent rel
ax e s towa rd
i ts equil ibrium
valuef
^
at a r atespe ci
fi
ed b
y a n a s s ociated
rel
a x at io n par am
eterAj
„
T h
ef
orm
of
th
e equilib
rium m
om
e nts/
f
^'',
tr ans
f
orm
at ionm
atrixM
q^m, s J i
d
inver s e tr ansf
orm
ati
o nm
atrix A/
*f
ollow d
'H
um
ier e s a nd
G
inzb
urg(
d
'Hum
ieres et al
,20 0 2
)
Th
is cor respo nd
s to solut io n of t he equ at io n:1 8
/
,(
X, +$
^,t + 1)
-
/
^
(
x.,i)
=E K
m^
r .{
r j
-L
)
+F
q,(
1 9)
w he r e
M
*^^
a r e c o e
ffi
cients of
t he tr ansf
o rm
at ionm
atrix inv e r s e and F
q is a c ontribut iond
ue toan extern a
l f
or ce g:T he c onstant re
f
e r e n c e de n si t y po is s et to u ni t y a nd
t he wei gh
ts wo = 1/
3, w ^ = 1/
18 fo rq = 1, ,6 an
d
Wq = 1/
36 f
o r g = 7, . ,1 8 Th
e s et of
r elaxat ion pa r am
ete r s a r e cho s en tom
inim
iz e t hed
ependenc e of
pe rm
e ab i l i t y on fluid visc o si t y(
Pan, Luo and M
i l le r, 2 0 0 6)
:A
j =A
j =A
g =A
io =A
j i =A
i 2 = -'^i s= -^
1 4=^
1 5= ~
i
(
2 1)
A
4 =A
g =A
g =A
i g =A
i7 =A
^g = a _\
' ' ^ ^^
w he re t he pa r a
m
ete r r >0 5
related
to t hek
inem
at ic vis c o si t y of
t he fluid:. =
^
(
. -i
)
(
23
)
T
od
e s c ribe no n-D
ar cy
fl
ow , a seque nc e of
ste ad
y-state ve
l
oci t yfi
eld
s are sim
ul
ated
w i thi
nthe gene r ate
d m
ed ia using spe cified value s of
Fo tod
rive t hefl
ow in th
e xd
ir e ct ionO
nc e aste a
d
y state vel
oci t yfi
eld h
asb
e e n re a ch
ed f
o r a g iv e n vah
i e of F
o, th
em
acr o s cop icfl
ow v el
o ci t yv " *
i
s c alculated
using ad
ensi
t y-w
ei g h ted
v olum
e av er age of
th
e po r e-s cal
e veloc i t yfi
el d u(
x i)
integrate
d
o v e r t hefl
owd
om
ainQ
,n >:r
od
r~
^
^ ' "/ n- n.'.P -
^
•'T he
m
a c r o.s c op ic v elo city is u s ed to c om
pu te theR
eynolds n um b
e r a sd
efi
n ed
b y E q 2O
nc et
h
e s equen c e of
Fo v al
ue s a nd
th
eir a s.so ciated R
eyn old
s n um b
ersh
a veb
e e n c om
puted
,E
q4
isuse
d
to c alc ulate th
e pe rm
eab
i l i t y and
ine rt ial
c o ef fi
cients a s s o ciated w i th
th
e anal
yz edm
ed iaR
e su
l
t
sa n
d
D
i
s cu
s si
o n
N
on-D
ar cianfl
ow simul
at io ns we r e perf
orm
ed b
y ru nning th
el
at t ic eB
ol
tzm
a n n sim
ul
ator o nthe Tops ai l co
m
put ing system
, a L in uxb
a s ed
clust er owned and
oper ated b y t he Univ e rsi t y of
No rt h
C
a r oli
na T he cl
uste r is c om
posed
of 52
0
c om
put ing nod
es, ea ch
equi
p ped w
i th 2
Intel
qu a
d
-c o re proce ssor s
(
M
od
el B
534
5
/
C l
o v ertown)
ru n ning at 23 G H
z a nd
12G B
of m
em
o ryT
ore
d
u ce sim
ul
at io n limes,t
h
em
od
el
wa s r a n in pa ral l
el
o n6
4 Tops ailcore s usingM
e ss ageP
as sageInter
f
a c e(
M P I
)
Ta
bl
e 1:G
ene r ated M
ed iaLognormal D istri bu ted Va riance
, a ^ P
oro sit y Range
T
'0 38-0 60
0 1 036-0 5 4
0 2 0 3 2.04 8
0 3 0 30-0 42
v a r
i
a nce s, a^
,
0
,0
1,0 2
, a nd 0 3
o v er a ra nge of
poro si t ies, se e Tab
le 1W
ith
in ea ch
po rosi t yrange, a sp he re pa cl^ing wa s ge ne rate
d f
o r eve ry inc rem
ent of 0.
0 1 To a ch ie v e stab
i l i t y,e a chof
th
epa clcings ha
d m
e an coo rd
inat ion n um b
e r s > 6 To ens u r e th
e e vol
ut ion of
a po ro us c ont in u um
,eac
h
porousm
ed
ia system
u sed
w i t h in t he sim
ula tio nsw a s am
ic ros c ale
repr e s entat ive ele
m
en t aryv ohi
m
e(
R E V
)
of
am
a c ros cop ic system
D
e vel
op
m
e nt
of
R
ep
r e s ent
at i
veE l
em
e nt
ary
V
ol
um
e sT h
eR
EV h
a s tob
e la rge e n o ugh
to c ap ture t hem
a c ros cale pli
ysics ofc onc e rn ,the r e
fo r e
th
enon-
D
a r cy curve
m
ustb
e ind
epe nd
e nt of b
oth
th
el
attice siz e of
th
e sim
ul
ated d
om
ai
n a nd
th
epo re siz e o
f
th
e gene ra t ed m
ed
iaT
ode v
el
op a por ousm
ed
ium
th
a t y ield
ed
grid
-i
nd
e pe nd
e ntr esu
l
ts, a system
c ontaininga set n um
ber of
sph
e res wa s pack
ed and
th
el
at t ice siz e wa s in c reased
u nt i l t
h
e non-Dar cy cur v e c onverged to o ne
d
efi
ni t ive solut io n To e xte nd
t h is system
to ad
om
ainin
de
pend
entof
po re str u cture,t he n um
be r of
sp he res wa s inc rem
e ntal l y inc r e a s ed
unt i lo nce againa
d
efi
ni t iv e solu tionf
or th
e n o n-D ar cy cur ve wa s re ached T he r e
l
at ive intrinsic pe rm
eab
il
i t y a nd
the in ert ial c o ef
fi
cients a r e sh
ow n a sf
u nctio ns of sp he r e num b
e rfo r
e a ch v a rianc e ofl
ogno rm
ald
istrib
uted
r ad
i i inF
i g lo
a
^
~0 5 0 0 I 15 0 0 2 0 0 0 2 5 0 0 3 0 O 0
N u m be rof S p he r e s
(
a)
Perme ab i l ity3 5 0 0 4 0 O 0 4 5 0 0
MS-'
V
-..
2 0 0 0 2 5 00
Num be rof S p heres
(
b)
In ert ial4 0 00 4 50 0
F
i gure 1:C
oeffi
cie n t v al
ue s r elat ive to the
as s o ciated R
EV
value s ob
tain ed f
o r t he perm
e ab
il
i t ya n
d
in ert ial par am
eter sfo r
a ra nge of
sp he r e pa ck i
ngsT
abl
e 2 o ut l in es th
e n um b
er of
sph
er es a nd
c ub
icl
at t ic e siz e s th
at were used
to gen erate aR E V
po r o u s
m
ed ium
for e a ch v a ria nc e of t hel
og-n orm
al l yd
istri buted r ad
i i co nsid
ered
w i th
in t heD
e ri
ved C
or r el
at i
o nsP
e rm
e ab il i
t yC
or relati
o nE
x ist in g c o r rel
at ionsf
o r th
ed
im
e nsionl
e ss pe rm
eab il
i t y were com
par ed
to resul
ts ob
tain ed
b
a s ed
o n sim
ul
at ions perf
orm
ed
using th
el
at t ic eB
ol
tzm
a n nm
eth
od
P
oi
nts ob
tain ed f
rom
si
m
ul
at iond
id
n ot ad
equatel
yfi
t a ny of
th
ef
u nct io n al f
orm
sd
esc rib
ei
n th
eb
ack
gro u nd
se cti
o nf
or t hef
ul l
r angeof
po r o si t y v al
u e s c onsi de red
F i g 2 shows how t he Ergun r elat ion de v iated si gni fic an tl y fr o
m
the sim
ula tion resul ts for thef
ul lrange of
po r osi te s,e spe cial
l
y at h i gh
e r por o si t y v alu es{
> 0 4 2
)
Th
eE
rgu n r elat io nd
eviat io nwa s qua
l
it at iv el yb
a s ed on a str ai gh
t tub
e ge om
etry of
th
e po r e spa ce and h
a sb
e enf
ou nd
tob
eval i
d
o nl y in a ra nge of R
eyn old
s num b
ers, 0 <R
e <7
5(
Ergu n, 1 95
2 ;M
acd
onal
dF
, 1 9 7 9)
B
ased
o n th
e pack m
g param
eter s s et b y th
eR
EV
c al
culat ions(
T
ab
le2
)
, t he ra nge of R
eynold
sn u
m b
e r s ach ie v ed
w i th
in th
e sim
ulat iond
iff
e r sf
or e a ch
va ria n c e Th
eR
eynold
s n um b
er ra ngeis
0
<R
e < 1 40 f
or t heh
om
ogen ous pack
ing(
c r^ = 0)
, 0< Re < 1 3 0
f
o r c r^ = 0 1,0
<R
e < 1 1 5f
o r a^ = 0 2, and 0
<R
e < 9 5f
o r a^ = 0 3
A
s t he
R
eyn ol d
s n um
ber r ang
ef
or ea ch v arianceincr eas es pa sse
d
th
e upp
erl
im
i t of
th
e Ergu n relat ion, th
e d iff
er enceb
et we e n th
e pe rm
e ab il
i t y v aluesf
rom
th
e sim
ul
ati
o n a nd
th
ose pred
icted b
yE
rgu n inc re as e sSi mulated Data a nd Exist ing Co r r elat io n s :Pe rme ab il it y
E
b
£rgu n Ca r m a nKo z e ny R um ptOup t B
——pa n e ta l . ^,0 A 0^=01 O o^ 'OZ
.
^
^
'^
^
^
Po r o sit y
F
i gure 2:C
om
pa ris onb
et we en th
e sim
ul
ated d
ata a nd
th
e e xist ing em
pi
ric al
r el
at io nsP
a n et alb
as ed
t hei
r perm
eab
il
i t y r el
at io n o n a set of fl
ow sim
ul
at io nsw
i th
a porosi t y ra ngeo
f
0 33
< e < to 0 4 5(
P
an, H i l pe rt andM
i l ler, 20
0 1)
, cor r elat ing to t he range of po r o si t ie s t heTa
b
le 2:R
epre sentat ive El
em
e ntaryV
olum
e sf
orG
e nerated M
ed
iaVa ria nc e
,ij^ Numbe r of S p he r e s
, jV^ Lat t ic e Dime nsio ns n ^ M
e a n p ix els pe r grain d iamete r ,P
"
(i 15 0 0 3 *5' "
3 2 8 4
0 1 150 0 3 8 0'
3 0 0 2 3 00 0 460'
27 9 7 7
re
l
at ion c o r respond
s to t he sim
ulated d
ata in F i g 2A
t the po r o si t ieso utsi de t he e xperim
e n tall
ysup po rted r ange, t he Pa n et a
l
rel
at ion und
e r e stim
at es t hem
e a sur ed penneab
i l i t iesT he
C
a rm
en-Ko z eny r elat io n
fi
t wel
l to a r ange of
th
em
eas u red
perm
e ab
i l i t ie s, bu tf
o rl
owerpo r osi t
i
es(
< 0 42
)
th
e rel
at io n starts to ve er of f
sig
n ifi
c an tl
yf
r om
t hed
ata(
F
i g 2)
T h
el
owe r po r o si t yd
eviat io n s ar e con sista nt w i t h th
efi
nd
ings of
oth
e r s(
P
an,H
i l pert a nd
M
il
le r ,2
0
0 1)
a nd h
aveb
e en at trib
uted
to th
eC
arm
en-K
oz enzy re
l
at ion o nl yb
eing v al
id f
orl
am
ina rfl
ow(
Prieur Du Pl
e s sis and M
asl i yah, 1 9 9 1)
,a n id
e a sup po rted b y the ine rtial pa r am
ete r r e sul tsf
rom
th
e sim
ulat iond
ataF
i g 4 showsh
ow t he inert ial cor rect ion to D a r cy'
s
l
aw, c o effi
c ienth
,f
o r t he sim
ul
ated fl
ow in cr ea s e s a s porosi t yd
e cr e a sesT h
ed
eviat ionf
r om
Darcy
fl
ow at th
el
ow
e r porosi t y ra ng
e sub
s equ e ntl
yd
e cr eases th
e v al
id
i t y of
t heC
arm
en-K
o z e ny re
l
at io n.T
h
eR
um
pf
-G
up te re
l
ati
o nd
eviates th
em
ostf
rom
th
e sim
ul
ated d
ata, overpred i
ct ing th
ed
a taf
or po r o si t ie s >3
8 and
u nd
e rpred
ict ingf
or po rosi t ie s < 38
Th
ed
ataf
orR
um
pf
-G
up te
is
b
a s ed
o n sp he r e pack ings w i th
r elat iv e stand
ard d
e viati
ons of
t he sp here-siz e
d
istri but ion,C
T o, of 0 0 9
4
5, 0 3 2, and 03
27
o v e r a w i de r ange of
por o si t y(
03
6 6 < e < 0 6 4)
and
R eyn ol dsn u
m
ber[
0 <R
e < 1 0 0) (
M
acd
on al d F , 1 9 79
)
T h
e ct^ = 0 3
d
ata w i th
in 0 36
< e < 0 4 2 is th
eo n
l
y sim
ul
at io nd
ata th
atfi
ts w i thi
n th
e e xp
erim
e ntall
y s up ported
reg
im
e of
th
eR
um
pf
-G
up te r e
l
at io nF i
g 2 sh
ow sh
ow th
ed
ata w i th i
n th
eR
um
pf
-G
up te e xpe ri
m
e nts co r r el
ates to th
eporosi t y ra nge w
h
ered
e viat io n s a re am
inim
um
, a n ob
ser vat io n sup ported b
y(
Pa n ,H
il
pe rt a nd
M
il l
er, 200
1)
A
n exponent ial fi
t of
th
ef
ull
r ange of
sim
ul
at iond
ata y ield
ed m
or e s at isf
a cto ry agreem
e nt w i th
th
e sim
ul
ated d
ata Th
e a ss ociated f
u n ct ion al f
o rm
is :oi
(
e)
= ae'
'''
,
(
2 5)
w
h
er e th
e be st-fi
tc o e
fi
i cie nt val
ue s ar e a — 19
45
x 10
^ a nd
7 =9 8
46
P
i g3
com
pares th
epr opose
d
co r r el
at io nm
od
el
to th
e sim
ul
ated d
atau o cc
Pr ed icted D ime ns io nle s s Pe rme abi lity
I
ne rti
al C
o r r el
ati
o nS
im
ulated
value s for t he ine rtial pa r am
ete r a r e show
n in c om
pa ris on w i t h v alue s pr ed
ictedb
yE
q 8 inF
i g 4N
ei th
er of
th
e c oeffi
cients pr ed
icted b
y Ergu n and M
a cd
o nald f
or th
eE
rgu nf
u nct ional fo r
m
(
E q8
)
prov id
e a s at isf
actoryfi
t to th
e sim
ula
t iond
ata, a nd b
oth
rel
at io n so verpr e
d
ict th
el
owe r range of
po rosi t ies(
< 0 40
)
and
und
e rpred
ict t heh
i g her ra nge of
po r o si t ie s(
>0
4 0)
As d is c u s s ed w ith in the la st s e ct io n
f
or t he pe rm
e ab i l i t y c o effi
cie nt, t he r ange of te stedR
eyn old
s n um b
er se em
s tob
e a sour c e of
er r o rb
et wee n th
e sim
ulated
da ta and
t ho se pred
icted
b
y th
eE
rgu n r el
at ion T heE
rgun rela ti
o n ha s al
s ob
e e nf
o u nd
tob
e val
id
onl y in poro si t iesra ng
i
ng 038
< e <0 47
(
H
ap pel
a nd
Br en n er, 19
65
)
,cor r el
at ing to th
e r ange of da
ta wh
er e the
Ergu n r elat ion sta rts to co n ve rge w i t
h
th
e sim
ul
ated d
ata befo r e
d
iv e rg ing at a po rosi t y of
03 5
Simul a ted Data and Exist ing Co r r elat io n s :In e rt ial Co ef f icie nt
^ 28
o
O 26
f
^
i:
S
«\
—
- Ergu nMa c d o n aldfl tal
■ o^ =a O
A 0^= 01
O 0 ^
=0 2
%
%
'
'
t
4
0 o;> S.
■ ^-» . *
Po r o sit y
F
i gur e 4:C
om
pa ris o n bet we e n t he si m
ulated d
ata a nd
th
e e xist ing em
p irical rel
at io nsT h
eM
a cdo
nald
et al
r el
ati
o n is th
eE
rgu n r el
at ion(
E
q 8)
w i th
am
od ifi
ed
B co ef fi cientbase
d
on th
e c om
parat iv e a n al
ysis of
n um
e r ous e xpe rim
ental
r e sul ts, includ
ingR
um
pf
-
G
up te
T
h
e anal ysis als
o to ok
in a c c o im
t dataf
rom
a w id
e r ange of
po r o si t ies(
0
1 2 3 < e <0
9 19
)
and
gr a n ular s
h
ape s and
siz e s w i th
t he go alof d
eriving anE
rgu n r ela
t io n th
at was ap p l icab
lef
o rpa c
k
ings of
no n-sp
h
e ric al gr ains(
M
a cd
onald
F , 1 9 7 9)
M
uch
of
t helo
wer po r o si t yd
ata(
e < 4 0)
ca
m
ef
rom
ex pe rim
e nts using irregul
ar sh
aped
ob
j
e cts, s a nd
and
gr a vel m
ixtures, a nd
a v ariet yo
f
u nd
is clo s e
d m
ate rials
S
inc e th
e sim
ul
ated d
ata o nl
y take s
in ac co u nt por ousm
ed
ia c om
posed
o
f
sm
o oth
sph
e res, ro ugh
n e ss co uld
e xpl
ain th
el
arged
e viati
o n sb
et we e n th
e in ert ial
param
ete rpre
d
icted b
y t heM
a cd
onal d et al
rel
at ion and
th
e sim
ulated d
ata at po r o si t ie s < 40
T
o provid
e ab
et terfi
tf
o r th
ef
ul l
ra nge of
sim
ul
ated d
ata, n o n-
l
in ea r
l
e ast squ ares wa s us edto fin
d
t heb
e stfi
t co effi
cients to a gene ral
ized
fo rm
of
th
e Ergim
r elat io n:(
2 6)
w
h
e r e th
e c o ef fi
cie nts ar eB
= 4 2 1, 71 = 1 58
and
72 =0
3 8F
i g5
com
pares th
e propos ed
co r re
l
at ionm
od
elto the sim
ul
ated d
ata■ 0^ =0 2
1 8 2 2 2 2 4 2 6 2 8 3 2 3 4 3 6 38
Pr ed icted In ert ial Pa r am e te r
F i gur e
5
: T he pr opo s ed
in ert ial
c o r r elat ion pred
icts t he sim
ulat io nd
ataC
o n
cl
u
si
o n
s1 T
h
ef
ull
s et of
sim
ulat iond
ataf
rom
th
e lat t ice-Bol
tz
m
a n nm
odel d
id
notfi
t a ny of
th
e exi
st ing pe rm
e ab
ifi
t y cor rel
at io ns,th
e Ergu n a nd R
um
pf
-
G
up te re
l
at ion sov e r pred
ictm
uch
of
t he por o si t y r ange w h i le th
e Pan et al and C
a rm
en-K
o z eny r elations fi t
d
ata w i th
in certain po rosi t y ra nge s I tw
a s ob
s e r v ed
th
at th
ed
e viat io ns of
th
e sim
ul
at io nd
ataf
r om
t
h
e Pa n et al
a nd R
um
pf
-G
up te r elat ions c o r r e spon
d
to t he r ange of po r o si t y v alue sl
ack ing e xpe rim
ental
sup port w he r e a s t hed i fT
ere nces w i th
th
eC
arm
e n-K
oze ny r e
l
ate to th
e inc re a s e of
th
e in ert ial param
eter I t was also noted
t hat a s th
eR
eyn old
s n um b
er incr ea s ed
pa st t he v al i
d
at ion r angef
o r t heE
rgun relat io n , th
ed
iff
e re nceb
et we e n th
em
e asu r ed
pe r
m
e ab
i l i t y and t heE
rgu n pr ed icted
pe rm
e ab i fi t y gr ewA
n e xpo nent ial
c o r relat ionm
odeld
epend
e nt o n por o si t y wa sd
e rived f
o r th
ed
ata s et2
T h
e in ert ial
pa r am
ete rsf
rom
th
efl
ow sim
ulat io nsfi
t th
e general
ized f
orm
of
th
eE
rgu nre
l
at ions but n ot a ny of t he e xist ing c o effi
cie nts proposed b
y Ergu n a nd
M
acd
on ald
A
gen era
l
iz ed f
orm
of E
rgu n's expre ssi
o n wa s c o nstru cted
to provid
e am
o r e sati
sf
a cto ryfi
t of
th
e sim
ulated d
ataR
ef
e r en
c e sBe a r,J 1 9 7 2 D yn a
m
ic s a}
F
luids inP
o r ousM
edi
aN
ew York; El
s evie rd
'H
um
ie r e s,D
, IG
inzb
urg
,M
Kr af
c zyk
, PL
al lem
and
a ndL
S
Luo 2 0 0 2 "M
ult i p l
e-r elax at ion-t i
m
el
at t iceB
olt
zm
a nnm
od
els in th
r eed
im
en sio n s " P h il
o s oph
ic al
Tr ans a ct ionso
f
th
eR
oyal S
o ciet y of
L
o nd
o nS
eri
esA
-M
ath
em
at ical P h
ysical
a nd E
ng ine e ringS
cienc e s3
6 0:437
- 4 5 1Ergu n,
S
19
5 2 "F
l
uid flow th
ro ugh
pack
ed
colum
ns "C h
em
ic al E
ng ine e ringP
r ogr e s s 4 8:89
-94
Fo ur a r,
M
, G. Flad i l la,R
. Leno rm
a nd
andC
M
oyne 2 0 0 4"
O
n t he non-l ine a r be
h
a vio r ofa la
m
in ar sing le-p has e
fl
ow t hro ug h t wo and
th
r e e-d
imensiona
l
po r ousm
ed
ia "A
dv anc e s inW
ate rR
e s o u r c es 27
(
6
)
:6 69
-677
G
e ertsm
a,J
1 974
"Est i
m
at ing th
e co efi i
cie ut of
in ert ial r esistan c einfl
ui dfl
ow th
r oug h po r ousm
ed
ia "S
o cP
etr ol
Eng p p 2 1 1-2 16
G
r ay,W
G
a ndC
TM
iller 2 006
"T
h
erm
od
ynam
ic al
l yC
o nstrain ed A
v e r ag ing T heoryA
p¬ pr o a chf
orM
od
eli
ng F low a nd T
VansportP
hen om
ena in Po ro u sM
ed
ium S
ystem
s :3 Si
ngl
e-F l
uid
-Ph
a se F
l
ow "A d
v anc e s inW
ate rR
e s ourc es 29
(
1 1)
:1 7 4 5-1 7 6 5H
ap pel
,J
a nd H
B
re nner 19 65
L
owR
eynol d
sN
um b
e rH
yd
r od
ynam
icsN J
:P
re nt ic e -h
all
,
E
ng lewo od C l
iff
sH a ss aniza
d
eh,S
M
a nd
W
G
G
r ay 1 987
"H
i gh
-V
elo ci t y
F
low in Por o u s-M
ed
ia " Tr ans po rt inP
o r ousM
edi
a 2(
6)
:52
1-53 1
H
a s s a nizad
eh,S M
a nd
W
R
G
r ay 1988
"
R
eply t
oC
om m
entsb
yB
ar ak
o n 'H
i gh V
el
o ci t yF l
ow
inP
o r ousM
ed
ia'b
yH
ass a nizeid
eh
a nd G
ray "T
r ansport inP
o r ou sM
edi
a3
(
3
)
:3
19
-32 1
M
a,H
and
DW
Rut h 1 9 9 4 "A
Num
e ric al-A
n al ysis o
f
t he Inte rf
a cial Dr ag For c ef
o r F luid-F l
ow inP
o r ous-M
ed ia " Tr anspo rt in P o r o u sM
ed
ia 1 7(
1)
:8 7
- 103
M
a cd
o n ald F
,E
l-
S
aye
d S
,M
owK
DuUi
en F 1 9 79
"F low t hro ug
h
Poro u sM
ed
ia - t heE
rgu nE
qu at ionR
ev isi ted
"I
nd E
ngC h
em
F
und
am
18
(
3
)
:199-208
M
cC l
ure,J E
,W
G
G
r ay a nd C
TM
ill
e r2 0
10
"B
eyo nd A
nisotrop y:E
x am
iningN
o n-D
a r cyF l
ow inA
sym m
etricP
or ousM
ed
ia "T Vansport in
P
o r ousM
edi
a84
(
2
)
:53 5
-5
48
Pan,
C
,L
-
S
L
uo a nd C
TM
il le r200 6
"A
n eval
uati
on of l
at t ic eB
olt
zm
a n n s chem
esf
o r po r ousm
ed
ium fl
ow sim
ulat ion "C
om
pute r s&
Fl
uids35
(
8-9)
:898
- 9 09
Pa n,
C
,M
H ilpert a nd C
TM
ille r 200
1 "P
or e-s cale rn od
ehng o
f
satu rated
perm
e ab
i l i t ie s in r a nd
om
sph
er e pa ck ings " Ph
ysic al R
e vie wE
64
(
6
)
:9
P
a nfi l
o v ,M
a nd M
F
o urar20
06
"P h
ysi
cal
spli t t i
ng of
n o nl
in ea r ef f
e ctsi
n h igh
-veloci t y sta
b l
ef l
ow thro ugh
po ro u sm
ed
ia "A d
v anc e s inW
ate rR
e s our ces 29
(
1)
:30
-4 1Pr
i
e u r DuP l
e s sis,J
a ndJ
HM
asl i yah
19
9 1 "F l
ow Th
ro ugh I
sotr op icG
r an ul
arP
o rosM
ed ia " Tr ansport in Po r ou sM
edi
a6
:2 07
-2 2 1R
um
pf
,H
, and A
R
G
up te 19 71"
E in
fl
i is s ed
er Po ro si tat u nd K
o r ngro s senv erteil
u ng im
W
idersta nd
sge s etzd
e rP
ore n str om
ung "Ch
em
ieI
nge nie u rT
ech
nik
43
:3
67
-3 75
.
Te
k
,
M R
19
5 7 "D
evelopm
ent of
a gene r aUz ed Da r cy equat ion " Tr ansA
IM
E 2 1 0:376
-37 7
W
ang, X , F Th
auvin and
K KM
ohant y 1999
"No n-Darc
y