Vous êtes sur la page 1sur 12

SPE 28792

Determination From Well Logs of Porosity and Permeability in a


Heterogeneous Reservoir
Jing Long Lin, Daqing Petroleum lnst., and H.A. Salisch, * U. of New South Wales
*SPE Member
Copyright1894, Soclely of Petroleum Engi.-a, lnc.
Thla papar was preparad 1or p!'M811tallon atlhe SPE Asia Paclflc 011 & Gas Conference hald In Melbourne, AUIIrlllla, 7-10 November 1894.
Thla waa oelecled for tatlon by an SPE Prograrn Comm- lollowlng revlew of lnformatlon conlalned In an -act eubmllled by !ha c,::anta of !ha =
aa =ted. have not by the Soclely of and are U,: =
any poeltlon o1 !ha Soclely ol Petroleum Englnoere, 118 olllcara, or llfl8l8 =-- Ulu8tr8tlone may
1101
be copled. The abelnlelllhould con1a1n conep1cuoua IICknowtedgmenl
TX 75083-3836, U.S.A. Telex 183245 SPEUT.
ABSTRACT
A tecbnique bas been develqm foc tbe petropbysical
of we11 logs mm litbologically amplex sandst<ne reservars
located in tbe Barrow Sub-basin in Western Australia. The
resems are sands of fairly high JXI'(Rty and low permeability
witb a significant amount of clay minerals, ammg tbem
glaucmite.
Principal ampment and cluster analysis were applied to classify
tbe electric facies associated witb litbofacies. Bayes disaiminant
analysis was used to ideotify severallitbofacies in tbe resems.
The different litbofacies are cbaracterised by distinct relatimsbips
between JXI'(Rty and aooustic intetval transit times as well as
between JXI'(Rty and permeability. 1bis paper presents
matbematical models foc eacb litbofacies ftr tbe detenninalim c1
JXI'(Rty and permeability fnm welllogs.
INTRODUCTION
Pocosity and permeability are impoctant parameters in tbe
evaluatim c1 resems. Resezvoir evaluatim genenilly indudes
their irlPntifiratil'!l, the det_Pnninsatioo of parameters
and tbe study of tbeir ecooooc significance. 1bis paper disaJsses
tbe methodology to detennine JXI'(Rties and permeabilities in
litbologically amplex sandstme resems in tbe Barrow Sub-
basin in Western Austtalia. The resems desaibed in tbis study
are of a amplex and widely varying geological and litbological
ampositim witb deptb. N'me stratigraphic la)U"S can be
Refemlces and illusttalims at end c1 paper
495
recognised mm are desaiptims, eacb la)U separated by
impermeable cartmatMemented sands. U>g cbaracteristics malee
it possible to ideotify five distinct litbofacies. 1be argillaalous
sandstmes cmtains, m average, abwt 48 percent clay, SOOle of
whidl, mainly glaucmite, was deposited as framewcrk grains.
Glaucmite, kaolinite and dllocite are tbe main clay minerals. The
distributim c1 tbe clay minerals in tbe resems occurs in tbree
main mnde.<: Slri.!Ch...l!"al , liuninsated smd ,
The objective c1 tbis study was to develq) a tecbnique foc tbe
evaluatim mm welllogs of beterogeneous resems of tbe type
desaibed above. 1bree key wells were selected as pilot wells. An
impoctant aiteria in tbeir selectim was tbe fact tbat tbey bad
reasmably amplete logging suites foc tbe J)I'OIX*ld intfl'pl'etatim
tecbnique. 1be wells wiU be called bere wells A. B and C.
The fdlowing logs were available ftr tbe tbree wells:
O Resistivity logs [Dual Laterologs]
o Smic logs [BHC]
O Fmnatim Density/Neutrm logs [LDTICNL]
O Spmtaneous Potential wrve
In tbis study tbe evaluatim c1 tbe beterogeneous resems fnm
well logs was based m grooping tbe resems into five distinct
litbofacies, based m log respmses and geOO>gical cbaracteristics.
They are, fnm tq) to bOOml:
2 DETERMINATION FROM WELL LOGS OF POROSITY AND PERMEABILITY IN A.. SPE 028792
q Bidmbatedsandymudstme
. Bi~ muddysandstone
. Bio@batedsandsttme
q Stllmgly~ted beds
Q slightly~ted beds
-
LxqonenL chstexaudB a)esdkximWm t Rnalysia
wereusedtixthedassificaticmandidentitkadmGfti~
kmiclogswreutilisedt oeadmate~ pausitybecause
theyafe theadytype ofpmositykWa_ forotheruells
drilledinthefidd. !kmicpmmitkw=tbm usedmtk~ty
~m~tie@*of ~ ~~
fhefive sdeuedlitbofadea aredifb=trn M* tkies.
lkequadmsd exiwdforeachl ithdhckahave~ite
tocunputethere@edpetrophysical~**@F.
Ihemethodology waslater appl.iedtGa _ well[mlllll,
outside of the@inalpikXarea.l13emauM wtreequdb
Cc43siatent withaxe analysisdata.
-ICTING LITHOFACm
Sipmanee
The Mennin@Ont)flithobde aisdfundmentalimpxtmcern
theevaluatkmofartxrvok, mcresowhen itis Meq@xnw
Lithe&ea andtheir architeUmearean@m CUItmlfiWkrrn
fluidflowmdfluiddistribudon intbepmes~tia~.
The lithologicalpammetm requiredfix log in@preWkm, sud3
asthetmtuosity fktm[a],thewxalled cfanentadmexpcaent
[mlandthe satumdm~t[n]mwk eadmatdas
aecumtdyRapoasible.Thesepamm@=smdo@y*mti
typesoflith*.
Cme&ima betwmamustic rntervdtmnait timeandpmsi?y
and betwem pmosityandpermeaWty vmiedti Mtitito
lithokies in this study.
tin-l nnn*r -w31vSC9A Flnm=rmdw zk - I * ~
A a~ ~- J - - J
t hedw audahof theresuwim lhe misdvity, smic and
_my@sbtiek* l& Band Guue first
standardhed tGminimke theeffeus ofscaks andunitsoflql
wiablesand thenrmnsfbned rntoanewselof~rnti
area Gf@nL5pal cunpments. Theti tburecnnpenents were
seleued fachlsteranalyakmle~~ -
ztitistudy.~-softidw analysis are
sbwn mthedmdrogcamufFigure 1.
Usingaaxrdation cod3kkntofabaIt 0.92aSabGUn*Wue
Rx theckWkamm Gfdeuric Meaintheae reauvokfiw
typeaofdectricfilcka wuedetmkl. Cmparedtidle*
thefivedectrictaciea cmespmdtGfil'edistinct MMildes
MhobgyandLo gRespmesofthe Five Typesof
Liihott&s
n3emavairswregfouped int(-)fivet)pesoflithohd eabasedon
VRllbganalysis.- rhegedqicd pmpcnksofthe fiwtypeaof
MA - I uk a., ?k e..Ahm.l me t Hl nw @
SW- WV U-J M S-VW=.
Facies1: BwturbatedSiwiy Mud@one
Facies2: BioturbatedMu@ SanaWme
Facies3: BwtudmtedSamir&ne
Facies4: Stnmgly C@bonate-CemntedBea3
Facies5: Slightly#rbonate-Canented Beds
Adetaikddex@don afeadItitciescanotbe@l.iskdatthiSdIE
The IMkrentlithotilckaaed istinui nthdrgeophydd pq=tiea
suchaseleUrical amustic andnudeardamUerM ea.GeGpllysical
wdllGgdRta eultainnmch Mxmatum al%nltgeophyaic!al ppexdes
c#lith&xka. Table lshGwthelGsreqmses tixthefiw typesof
Iithokies.
Table l- AverageLog Valueafathe
FiverypesafLMlokkY
1 1 2!3:4~5
.. . ......... . .. . ........... .................. . ...... ....... . ...... ..... ................ .......... ........... ...
R i 2.228! 2154 ! 2514 i 4397 i 4LB6
................ . ...... .................. .. .... ... .... . ............ .................. ..... .... . ...........
Rum!
2A39 ~ 2103 : 2179 : 12.476 ~ 3.642
+ .."-"" ...."+""_- ......u.+-- ......""".f"-"""..-_" ... . ..... . ... . .. . .. ....
*;
2359 ~2298:2207:2529i 2.304
.................. ................ ........... ............. ....... . ... ........ .. ..................... ..................
h
~ 116591 : 122673 ! 130.34 : sa.055 i 105.77
... ............ ... ........... ... .. ............. ... .. ...... .... . ........ .. . . .... . ..... ... .. . . ... .......
GR ~ 103.245: 112531 j loa.275 ~ a6.34 ~ 91.512
Wherel?t-deepinveatigati mreaistivity
n -: --11.. & . #Ml 4.* .4.,
~ - Uuut mpl nal uuuy UA.imKu l c al auvuy
pb-btdkdmsity
A(-amustictmnsittime
GR-nat uml gammar ay
IhediOkmma inlc)gmspmse afathefive typesoflithdadea
canbeusedtGidutti@ t hel i t hof adea.
I&ntMWkm
Bays Wdminan tanalpisw asappliedi nordertoattenpt tG
. . .
teammgt hefive~e frocks ntempmsea fed
496
SPE028792 J.i.,. LiN and I-M. SALRCH
~
fim the (3amma-Ray,Hv@ md sdc W - M fm
thispulposc
Oneofthekey steps indkdndmmt anaUsiaist ockosethe
smndafd aampkswhicb eanrepresmt acezlain litM%cieato
estahbhthedbhninm tfunetkm withlnaeacemaey. m
sumdardsatnplcawaeseleued bast!donthereadt$ Ofclusta
analysisandeaedata
Bays dkxbinm t fiutcdon Fix) is daiwd h BayEs
v
Fi (x ) = iqgi j + G + ~CUzk
t d
H, z ....0,
wherez
x. (XlXz,...,Xv\ a stqie fixfidiKZ@kIM~~
V.nukle$ofvariabk
G numb ofcategaiea (G5)
q- nJN
N-tad numberof samples
nr td numberof theith categorysamples
6i*Qr~t-
. . .
DMnmmantfktcrsvmet hendmmmed. Iheyafeexpmsed
inmatrixftxl nasfollows
Inthedkahnhn t - a -Pk x = (h h .. .. d h
. . .
kxmunant analysis isentaed into the above dkhnbnt
funcdals toCaleldatefunaien wllues.
If F{x) = MAX{F~x))
j=l, Z..., G
.&t@2 X bekmgstothe ith eategay.
Iheearectdkxhnbm tradofa standardsampkaisuptoso%
fakieslandfacies2 90%fabcies4and ltM%fahcies3
andtiKiea5.
T- identiikadm ofiitho@cksk2r-*@i-welisAIi atidc
was dcae using Bays dkiminm t aoalysis tdmique. fhe
euIeudkiminm tmtioiscketo 84.3%fix WEUJ%81.5%
fa wellBand 78.7%fa U@C.
Paoaity isanhnpatant paramem in the emhmticaofa
reserv& Theamount ofhydmcmbm[cil andkrgas]ina
reaavoh isesdnMt@ amcflgahathing$ ffompaeaityvaluea
and fianthe thiclmessandarealextensim ofamsemoir. Total
Paosityisdeiiaed astheratioof thetotalvoidspxe intherock
toitslndk volutne. Eff@ive Paosity istheratio of the
interconnectedwuidspaceintherocktoitsbul kvohnne.
IWeewdls ~Band Cwheze~, CNLand BHCSeaiilogs
had bee$ltakenwerewlected askeywe'lls. saniebgsareused
hereto6tknatewfmaadon po@tyofthegiven. KXrvo&sbecause
theyarethe adylypeofpaoaity logsavaila bkinallthevdls
drilledinthekld.
tmtiminter@. Tkamount ofclay minemls inamsavoiris
usuallyeadms@dfiun Giun3naRaytmd spontaneousPotential
logs. Theseesdmatmdo na namally rqxesent the true amount
ofdayminaals duetodiff&encea inthenatuce ofclayminaals
inthesands andinthe al@ent shaka.Ihe _toccumence
of2WioaaiveminemlssuchaspotasaiumWdspar,glauconiteand
ahaheavy minemls inthereamoim studied haedd to the
diffiadty. lheshale distributicminthemsexwir rn its three
modtxstmetm& laminar anddkpmed. makea itditlkultto
eadmatepaosity aecumtdy with traditional*suc12 as the
Wyllietimeaverageequadon afk emec&on fa elaycontent.
Topredictporosity intheseresemks fmlnlbdl logs themain
prcMems to besokdlK etheeffkmofcarbma@ and clays
whichfampartofthercck aswllastheefikc tsofthedifkrent
Iibfkiea.
497
4 DETERMINATION FROM WELL LOOS OF POROSHY AND PERMEABILITY IN A..
SPE 028792
Atechnique has beendevelqedh exetoe admate porosityand
~ty mheterogeneous
msmmirsof dlisnatme.
Fiveidmdfiable oflilhohcie swredetumid fiunwell logs.
Distinctrelationshipsbetweenfennadcm@ty and S(XliC ha%d
timeexist tiXeach lithotkciealT@ure2].Ik five
matkmadcal
models fmthedekzmhm mofpausity fian-tnmsit
timmwremtabhshed bystatisdcaltedmiques.
cIlM410tAnal yai a
Fcrthestudy ofpcrosity, 306samples wereusedikxn thethree
wlls~Band C.
c-me pcrositiea and XOtlStiC transit times - aoss-@tted in
figure2. Tldsfigureshowsthataud atapointsa rekXatedhtwo
distinct areas. (lneofthe miscanpmedo fsandstcmeaudshale
eknent& the OthK Culsist mainly of ~ted
cmprments. Figure 2 also indicatesthat distinu relationships
exist betwtxmpau6ityand acoustictransit dmetithefivetypea
OfIithofacks.
St udy of a ModeJ
lleobjective ofthisstudy wmtotrytofind better sadc
transfamswith thesmalleststandard umcfcrthed@mnhmdcm
cfporosity. ThX@ewell known porosity Uanstlmnswtm
. . .-. ...
cumukrd me wyme rimeawrage equab, me Raiga-
@ZllUlfXZiU WkOUStiC-CXI=tfX equatkmandthe
Raymer-Hunt-Gardnex equatkm.
Based mthethreeequaticxw threemodek linear, mcipmmland
lqarithmicwreused todekxminesadcpaositytransferms.
A Iinearmodd fkxntheWyllieequationk
@
,=a+b At ....................(2)
A re@mcal modelWed on the Rayner-Hunt-Garduerequatkm
Canheexpmsedw
@,=a+-&
..................(3)
A logarithmicmodelduived tlem the AcrosticFczmatim=
equatial k
log(l -$)= a+b.log(&)..........(4)
'lllethreemodels wemtalominmgressial fmt?ad3lithofa&s.
C4mhtim andenuanalysis wreusedin thesmdy ofthe
*. Cmdatlm. analysis isamarhematicdmelbodf crthe
Sh3dyofcmdatml betweenvariabk Twomriablm ammae
lhl=ly cmeMabk
theclmez thevalue of thecmelaticm
coel&imtlJqis tolcx-1.
Standarderrs [SE]is givenas follows.
i=
~T,-Tc, Y
SE= = ~ ~
.......... (5)
where
N - numlxxofsamplcs
fG -- [cOKS] valueof the ith sample
Ti- logderivedvalueof theith sample
llueefmmsofpresentatial werecmsidmd fa tic trawl time!
tircmekaca. and fcrerrm analysis.l%eresuksofcmelatim
analysis anderrcr analysis aregiven in Tables 2and3,
WP@~Y.
Table2ComeMkm CodE&nt farthethreeModels
:I%icu \FaLicd :Frlciea :Facics ~Fadcs
jl
j2 !3 :4 IS
"""." . . ... . ... .... . .. ... ...... .. ... .... . .. . ... . . ... . .. . ... . ..... . .......
* ! a74n i 0.6635 ~ 0.7620 ! 0.s052 : 0.74s9
. . ... . ... ... .. . . . . ... . . . ... .... .. . ........ . . . . . . .. ... .. . . . . ... .. .... .... .....
-
: 0.746s : 0.6614 : 0.7574 ! 0.8065 : 0.7603
bar-ic!!ii;
0.7471 oti2a 0.7609 0.s053 0.7547
Tahle3-Stauiard Errarforthetbree Models
j Facies ~Fa&s/i%i?d jl%ird jFuiEs
:1 ~z :3 :4:5
. . . . . . . . .....~. ... ........+ . ... ........ ...p . ...... . ..... . ...... ... . . ...."......
I
* !1.399 i 1.622 ~ 0.965 ! 1.S32 ~ 1.s64
.-.-.. "..-. --&-- ....--4 .-- ....--. &...-...-.-". !..-" _-A.._.""."._ I
- :1~ il@
i 0.969 ; 1.SW i 1s43
.. ..... . . . . . .... ... ... ......... . . .... .......... .. . ........ ...... .. . .. . .... . . . .. . .....
~
: 1.399 ! 1.623 j 0.%7 : 1.s11 ! 1.534
. .. .. . . .. .. ... . ..... . .. . ....... . ... . ........ ....... ...... .. ... . . . . . . . . . .. ....I
Thereadtsoftheandysis mabks2and3] indicaethatrheliuear
models t& fhciesl, bciea2andfkcies3 havethehigheat
cmelath melkients andtheloweat srandardemm, =the
mc@mcal*@ethebeatmsuhsf m*4andhcies5.
498
SPE028792 J.L. LIN and H.A. SALISCH 5
Basedmthermults ofcareladcmandaror andysisthel inear
models aodrecipmcal models wexesekedheretomtimate
famaticmpaosityfcr faciml, 2and3and for fiwks4md5,
-~Y.
l%e equatkmsfa axnputing paosity ffansatic travel times fer
thefivetypmoflithofkies byregressimanalysis=
Equations
-fwfwiesI
OS= -0.49503+ 0.162939At
ofa model f~ thetwosections.
Basedcmthe results of careladm and arm analysisthe linear
model wasseleued fcr Sectkmland thereciprocal model was
sekaed fa Sectb 2
Jnthe`One Group' modeJalinear equationwasusedto@imate
paosities fkxnsadcl ogsafkresukshad beeaobtained ina
manner similactotheTwo Group model fa correlationand
ara analysis.
TabIe5give8 thecamparison of the three models fcr the five
-fw$lcies2
@S= -0.29080 + 0.193642At
-fw@cil?s3
OS= -9.91978+ 0.294179At
-f-wjack 4
@= 46.68793-2704.26 I At
-Jmfmus 5
@S= 56.73825-3653.45i At
he valoes fixthe arelatim codiieieat IN and the standard
ara[SE] czdadatedbasedcmthe306sampksf rcmweJls~ B
and C by using the above five equatims are 0.885088 and
1.520843,respecdvely.
@np@sonoftheModds
Inthisprcje ctwveralmode lsw reusedtorelinerhe
cksitbdcm. The OneGroup model places the entire interval
into cmecategay theTwo Grapmodel does sOiuto two
categaieaaud the`Pive Group' model dividmthemsmmimhUo
fiveCategcrk?.
llecmeladcm coefficiemsa adstandadara ti3Msamplm
lkxnthethmewells menticmedabovearegivenbyTable4.
IntheTwo Groupmodel themaavoir Wasdivided intotwo
parts calledSeuimland Seuion2.1k fama~ds@
ihciml, kies2and kiea3withsands aod shakss@.km2
~
&ies4aud5witb ~ted ekueats.
fhetbree *afacou8dc interval tmnait time: linear,
~mdl~-~ .
conmkedrnthe sektitm
Tabk S- CampwkmafStandaml Error [SE]
oncGmup / 3.8688~ 1.6492; 3.0133~ 2.0619j 3.260s
mold:
............... ........ . ................. .................... ..........
TwoGraqI : 2.S6S6: 1.85% ; 2.1648~ 1.850S: 1.7503
lldd~
............................................ ........................................ ......................................
FiveCuwkP~ 1.3991j 1.6219~ 0.%54 ; 1.8133i 1.5642
MYkl:
Applkations
l%el%eG .rmpmodelw asusedfm thedeknninationof
~titi~rntie k@mUs&Bmd CmdtiafdmU,
D.'Iheresults franthe `PiveGraup'madel werecanparedwith
thoseb a modei usedat presmi til WAfield. F@-. 3 and 4
givethetworeauksfcrwellBandwellD,m pectively.
h the preamtlyusedmodelpaoaity is dekmninedby the Wyllie
timeavemge equatkmThe apparent sadc matrix valoes [AM
and apparent satic fluid values[Ad am obtainedby singlewell
regreskms faeachofthe wdlsdrilledin the field. 1kyvary
llcm well to well. he pafametas fa the five wellsmentioned
above aregivenrn Table6.
mlbie6-sdc Mairixlmlimki vaiuI%hseii eiijhime
Wdt~A!B~ C~D
............................ . ............. . . ......................................
Ab@c!ua)~ n i 61 ! n i 73
...........................~~
&@Mcda) ~ 302 : 310: 2s5; 306
l%e Iithofacks fathetbee keyvdswaeidmtified by the
IlqK!s didlnbl tfimeticms which h@ been @abMhed and
wereshown earlier inthepapa. lk paositieawaethezl
mldated Ikxnkweltolevel byusingthe appropriateequations
fbr thediikxent lithokies.
Figum3givesthetwormults fawdl B.l'kfigureindicates that
6 DETERMINATION FROM WELL LOGS OF POROSITY AND PERMEABILITY IN A.. SPE 028792
the pcmity values derived tkxn rhe We Group model are
ti@~mdcl~m tie-ti, _yktieti_ti
85t093me&es and from 103tol10me&ea. Onlythelastdigits
of the depth figureshavebeendisplayed.
llgure 4 showsthereds for wellD. The figureindieateathat the
values of porosityekmated tiwn the We Group model are
againhigberandclosa totheccmda@ eape&llyin therntervals
fkom102metres to107metres andtianl14 metrestol18
metres. only the lastdigitsof rhedepthfiguresare shown.
Thecmelation inefficient and standarderrer valuesof the four
well showthat rheresultsffomthe FNeGroupmodel are mere
accuratethanthosepreaentlyused.Detailsmgiveain Table7.
WeU ; A! ii: c:
~
..................... ............................................. ..................... .
R~ ! 0.704 : 0.768 ! 0.557 ! 0.831
..................... .......................... ...............................................
R [5] : 0.715 : 0.823 ~ 0.654 ~ 0.854
..................... .................................................. ...............................................
sE~ ! 3.247 ! 2.905 ~ 3.015 i 2.114
..................... .................................................. ...............................................
SE[51 ; 2521 ; 2.170 ; 2.529 ! 1.953
[ p] - Pf Wdy U8ed I UOdd
[ 5] - FiveGroup model
Permeability ofarock ismeof the most important flow
pammetm asdated with suktrface produaion and injectkm
opemtims. TheimpaTance of theparameter isrdteetdbythe
numbez of sources -well logs mea rmd well testing ameng
** t@l & employed to ~ ~
ty. well log
evaluation has the special advantage of ecmany and tbe
cxminuity of measurementinthe esdmate ofthiiparameler.
Permeabilityofafcrmarim @intluX averycanplex timaion
Whiehisaffectedbyxrumyfactmssuchas porosityandptre$pa&
&mWeristic&~ amamt and diatributkmof Ay !nimm@
rockmatrix canpositicn andsizeofmaaix grains. lhe
cmmdlingfaetm fcxtheva lueofpemneabilityin arewmoirare
dependent en its geologicaland sedimentarypmpades. Many
empirical cmelatims have beelldevdoped toeadmate
penneabiIityw, atleasLapameaMity indextiwdllogs.
lheinterval studiedisdmaamsd by its mnplex Iithdogicd
cunpositicmand largeamamt of clayminti by Iigh pmosity
and low permeaWty. Based (m these heterogexmous geological
chameMMcs atedmique has been devdopedhtxefera
ealdnuals eadmateofpermeaWty * welllogs.
Thememoirs showfivetypesof tithotkciesidenti6ableMm well
logs. There is a distina relatkmship between pmosity and
permeabilityfcc each lithofaciea.Ihe empirical cometadcnsfcr
thetypesoflithofaeiesweremtabikhedbyregreaskm.
Cro6s-Plot Analysis
Thepermeabilityof the mervohs wasstudied using306samplea
franwells AjBand C.
Cm pcrosity and are ~ty wae cross-pklttedin I@ore
5. Itindicatea thatthe datapoints are spread over two areas.
Thosemone areaare eunpoaedof sandskmeand shaleeIeanent&
theothers emsistsmairdyof mbMXXmmted Cunponalts.
Furthmmq figure 5 showsthattherearedistinadatknships
!x%w&l!pc=oskyandpe mneabdityforeaeb Ofthefive typesof
lithofhcies. l%ismeans that pemeaMity is not only dated to
PmOsityhtalsotothe intrinsicchmetaMca Ofeachlithofacies
he 6rst famula relating measurable rock properties with
pmneabilitywasproposedin 1927by J. Kozmy and modifiedby
P. c. Carmam
a+
=ss;(l-d
................. (6)
where
k-~ty
O-p(xoaity
St-grain _areaperunit8did volume
If thelithdogy ofafamatim deea not vary, the total grain
surfaee [SJinthe fcmmdama ybecnnsidued to be constant.
ll~pmeabdity vdllbea fimaiat of porositytbradistina
I.ithobcies.
The next majx S@ was taken by Timur who extendedWyllie
and Rmeseqiri&mbasedo nMomtory studies of 155
sandstonearea. lle Timurequatkmis describedas follows:
k=O.W$
........................
(7)
k - ptmrteability[red]
@- p(xosity[%]
L-imedueiblesammtnm [%]
Inedueible water satumticm iaaiimaicm a- the rock
Ckma+aeri. E the Mhok)gydoes not Vagy, tbe .irbk
500
SPE028792 J.L. LIN and H.A. SALISCH 7
-Watef Satm-ath SkM remem?Qm!tm -has W@
pumeability wi.llbatidmof@tY fcradisdnct
lithokka.
'IEereareatherequations similartothat@W=bY ThnWsuCJIm
theMinis and Biggsequation.Theseequatkmsam beqressed
witha generalfernmhx
k=c~
WI............................. (8)
Areaemir withvariationsin lithology, with diffemneeaeither in
the canpositicm,stmciumw Oth= ~
m dqlent3y
becksitied into severaldiserea Iithoiheiea. l%epmneabilityfcr
agivenlithofacies may bedeaaibed~ yasafuncticm
of the ptxaaity. lle gceatezthe number cf Iithotkiea iuto which
amervoir istroken up, themmeaccmate will be the distinct
relationshipsbetweenpermeabilityandparosity .Ontheothez
side thegreaterthe numberof lithofiteiesthat are demmimd fer
agiveninterval theharder willbe their identificatkmby
. . .
dmmmmant analysisfian welllogs.
he merwirswexegroupedintofivetypesoflithofaciesbasedon
both their lithdogical properties and rhe pmsikd.ityof their
identificationffan well10gS.
Twomodelswvremnsideredin the studyofpemneability
log(k) = a+ b.log(0).m$...o....... (9)
log(k) =a+ ho ............. (10)
Thetwomodels wereentefedintoregmasicmfmeashlhltofSciea.
fheccrrdatic manderrc ranalysiswereused toselectthetinal
modelfff theesdmateof pumeahilityfhxn well logs. l%eresults
Ofcmelaam anderrcr analysisare givvnin Tablea8and9,
WAY.
Thecameladm me&iemsinTable8 and standard errcxsin
Tabie9 wre calculatedfm the i-t of ~ma.-
- flQg&J] and logarithm of ~
riea 6xlm logs
@og(&)]. lld.smeans thatthe wIhKarnbUhTw 8and9
reQecttherelaticmshipsbMveenlog&)andlog&).
?We S- ~Qweiat!t! Qdfide@ fmtheTwo Models
jl=ui?a ~Fades ~Fu5ca /F&ied ~F&iea
~1~ 2~3 ;4
!5
................... ........ . ............ . ............. ........ ......... .........
II@)= * : 0.7166 ~ 0.8937 : 0.72s6 ! o.75m i 0.9114
................ ........ ...................... .................... ..................................... .. ............... ...
w)=
~ 0.7209 ~ 0.8983 : 0.7302 ~ 0.7438 ~ 0.9046
**) i i
TatAe9-Standard Error forthehvoModels
The results of the analysisin Tables 8 and 9 indicate that the
semi- logarithmicmodels fix tkcies4 and f%eies5 have higher
emdatkm coeliiaent valuesand lower standarderrs, and that
thesame appliesto theduallo garithmicmodels fmikcieal,
fkeiea2andtkies3.
Equations
a~;ti~e@ W _ to obtain a mgreasicn. l%e
esdmateofpexmeabilit yfcanwel llogsfcxthe
fivetJQesoflitl@cies afe:
-farjhcies I
log(k) = -11.W + 7.732442log(cD)
-forjileies 2
log(k) = -16.5817+ 12.25489log(~)
-fotfmies 3
log(k) = -11.6038+8.88765log(Q)
-faPJacias4
log(k) =-2.68508+0.142257(0)
-fwjheies 5
log(k) = -2.07977+0.141827(@)
Themmelaticmeoeilicient~] and staudardmar [SE] ealadated
ftxthe306samplea fkotnwells~Band C byusingtheabove
fiw tt@tkXIS are 0.9530 and 0.2130 [b Iq ~) and 1~
~] ~d 0.9292and 2.69f31[h lcmre and&], mapectivdy.

.
Conma&onof the Ma@a
Inthis~several models wae used toohtaina~
,.
elambtm TheOne GroupmodeLtheTwo Group modeL
andtheFiw Gmupmoddt&whieh theinterval was divided
intoone twoand five&k& ~~y.
IntheTwo Group model thereaer@r was divided into two
parts calledSaimland Sectia12. llefimneris eunposedof
t%iesl, tt@es2and fkies3with sands and shale$thelattex
ccm@seahcks4 and5witb arbonab eunentedekznents.Tvm
mode& semi-logarirbmieand dual logarithmi~ wue studied
within theTwo Gnmpmodelfman eadmate ofpmneability
ffcmsonicpcrositiea.Basedonthereaultsofemela&m and~
analysis the dual logarithmic model showed the highest
501
cmelath me!Ticient and lmw?st standarderrcr in Secdon1; the
sameappliedtothe semi-logarithmicmodelin Sectioo2.
h theOneGroupmodel a semi-logarithmicequath was used
toesdmate pmmbility ~sadcpmosities acuxdiog to the
resultsof thecmelatioo and arwanalysis.
Table 10givesthe resultsobtainedfran the three modelsfcr the
fiveIithofa&s.
Table 10- Results horn Standard Error [SE] Analyais
k)rtbetbreeModelsi nthefiveLitMkks
~ Fadcsl ~ l%de.s2 ! Fuies3 ~ l%dcd4 ~ -5
.......................... .................... ........................................ .............. ..... .................
Olb?Grulp ~ 0.3433 ~ 0.2260 j 0.1937 ~ 0.3733 / 15797
-;
.........................+................... ....... ...........+_.._-_ ~
: 0.1740 Twokup ! 0.3018 ~ 0.2202 . : 0.37230~ 0.3222
IuxM
.......................... .................... ........................................ ....................
me GTc up ; 0.1987 ~ 0.2165 ~ 0.1515 : 0.3083 \ 0.1755
IIKddj
TheFheGroupmodelshowst obethemestaccumte model.
Table11makesacunparison of thetbreemodelsfw 306samples
fium weUsABand C.
Table ll-Resuits of Carreiation and Error Anaiysis
fromtbethree Models based ondatafrom306san@es
I : CM3rcW1WM: Tw@wPIIXIIM ! FivcGKNPdell
R; 0.7960 0.9413 : 0.9530
...................................................... .............................. . ........................
SE; 0.42s5 ; 0.2373 ~ 02130
I nthestudypsenteda firstlyI ithc4icies$wxeidmti&dfa
thefour wellsbytheBaysdisdmhmt functkms. .Pmsities
wethenadahtedusing theapprqmm equation faeachuf
thelithohcies. ThepemeaMitywasedmatedI kantheFive
Groupmodel.
R~6@mtiew~Kti@B. ~efi~rnti~ti
thepermeabilityvaluesderivedIkemthe FNeGroupmodelare
maeaecmate Thisismostappareat intheintemalsti 102to
107m. and fixxnl14to l18m. Onlythelastdigits of thedepth
figureshave&emdisplayed.
Figure7makestheampaism oftbetworesultstlr well D.
Alsohem theresults ticmtheFheGrmp model afemme
accumte. Again, cmlytbe lastdigits af the depth figmes are
Shown.
Table 12- Correlation Coeffklent end Standard
Error for the three Wells
A: B~C~D
.................. ............................ . ............ . .......... .......................
R~/ 0.576 ~ 0.663 j 0.597 ~ 0.623
.................. ............................ . . .................... . ............
R [5] j 0.611 ~ 0.713 : 0.694 j 0.716
.................. ........................... ........................................................
SE~ ~ 0.993 ~ 0.891 0.872 ~ 1.049
.................. ........................... ................ ..................................................................
SE[5] ; 0.776 ~ 0.619 ~ 0.685 \ 0.842
[P] - Presentlyused model
[5] - five Group model
HE isoittshmva he fiew Wmii-i giivminuaeamr-zi
pelmeawtyvzdues.
CONCLUSIOhQ
Applicdons
llel%e(kmp model wasusedfaan esdmateofpmability
inthethree pilotwellsA, Band Caodrnawell Doutsideoftbe
pilot area Theresults fkxntheFive Group model were
canpmedwiththowt kcsnthemetbod prmently inuse in this
field
h the melbod used at present a semi-logarithmicequarh is
applied todelemme theptmneabilities duivedfcr weUa~B
and C.
Theeqwkmfat hemefhalusedpesentlyix
log(k)= -5.4524 +0.23810
1. The Iithofacieaof the memoirs Subjeaof thisstudycanbe
identifiedffan well logs.
2. llediff&ent lith@wies of these mervuim have disthm
rdationsbipsbetween@ty and X!U3StiC intervaltraosittime.
3. ThedilWent litboikks of thesememoirs have distinct
reladonsldps betweenpaosityandpexmeaMty.
4. The techniquepesemedintbispsptx -heecalledl%e
Group model- ismtxe~than thepresmtly usedmodel
tirheevaluaticmoftberesmdrs fiomw?lllogs
5. Thenewmodel isrdiable andaccumre rntheevaluatkmof
heterogeneousresmmimftxthegiwm amditicasand rntheama
studied
6. Itwuddbe ofgeneral interest toapply tbemodeltoctber
rmavahsand testtheextmtof itsvalidity.
SPE028792 J.L. LIN and H.A. SALISCH
ACKNOWLEDGMENIfl
lheauthm wish toacknowledge tbeassistac eoftbecil
can@eswhichsuppliedth edatathatma dethisst udypossible.
Theytuealso gmtefulfcr rbecmpuatimreceivedattheCmtre
h PetroleumEngineeringof the Uniwxsityof NewSouthWales
througheutthecourseofthiswak
1.- JohnC. Rasmux A Summaryof the Effectsof VariousPore
Geanetries and Their WettaMity cm Masured and In-situ
Values of Cmnenraticmand Samahon ExpaIems. he Log
AMIYSLVol 28, NO.2 1987.
2.- 3). Maricm, A. Nur, aod F. Alaberc Modding the
R4atimships Belween sonic velocity, Pmosity,
~vv
and Sbdiness m SaucLSl@g aud Shaky Sand, SPWLA
TltirtielbAnnualLoggingsympdllm!1989.
3.- A. E. Bussian:Eleurical Cenductamzm a Proms Medium,
@@@c& VoL48, No. 9,1983.
4.- U. Ahme&S. F. Crary, and G. RCoates :
~v
Estimadcxuthe Various Sources and Iheir IIItemlationship,
SF%19604,1989.
5.- M. R WyUieand G. H Gardner :TheGmerahM Kozeay-
- Equatien,Wcxld(XLMarch-1958.
6.- J . Kumar:NewClint OffersFast PmrneabilityEsdmat&,
WorldGil , Fe&wry, 1971.
7.- GeorgeRCoatesand J. L. Dumanoic ANew Appmachto
Improved Log-derived RlmeaMi&, 3chhllnbexgtX well
!ltzvim - Texas.
8.- RichardH. Benneq KathleenM. Fischtx,Da- W-R
and RichardRezak Pwcmelryand Fahic of Marine Clay and
&bonate sedimmts Deets d RmneaMi~, Marine
Geology,P. 127-15? 1989.
9
9.- Lev Vemik and Atnas Nuc Pe&cpbysicalClassMcarkmof
SiIiciclasticsh Lithdogy and PmosityPredicdm b Seismic
Velocities,AAPGBulletin,V. 7Q No. 9, p,1295-1309, 1992.
503
10 DETERMINATION FROM WELL LOGS OF POROSITY AND PERMEABILITY IN A..
SPE 028792
FfGURE1 ELECTRIC FACIES FROM WELL LOGS
I Rlld I I RIM I I
Rud
1
1.
Fi GURE 2
EEi EEi l
Wells A, B and C
z 35
.
8
30
-25
E
20
A facies 4
g 15
I
Q 10
*
70 80 90 100 110 120 130 140
ACOUSTICINTERVALTJUUJSITTIMEpO&Rwedftl
504
SPE 028792
J.L. LIN and H.A. SALISCH 11
WELL B - COMPARISON OF RESULTS BETWEEN THE FIVE GROUP MODEL
FIGURE3
AND THE PRESENTLY USED MODEL
35
30
~ 25
z
20
g 15
g 10
5
0 F?
R=O76S SE=2.905for pr esenuy umnn-
R+l QO1 CC-9 17nf?w12hmGr mm model
80.00 85.00 90.00 95.00 loo.(x) 105.00 110.00 115.00 120.00
DEPTH [mares] - Oldylast digits of depth figures are ShOW31
WELL D - COMPARISON OF RESUL~ BETWEEN THE FIVE GROUP
FIGLJRE4
MODEL AND TEE PRESMWZY USED MODEL
40
E 30
E
20
8
M 10
2
0
100 105 110 115 120 125 130 135 140
DEPTH [mares] - only last digits of depth figures are shown
12 DETERMINATION FROM WELL LOGS OF POROSITY AND PERMEABILITY IN A..
SPE 028792
FIGURE5
STUDY OF PERMEABILITY
Wells A, B and C
- 100
1!
g lo-
&,
G
/
~ 01
n
s
4
~ 0.01
, # ,
0 5 10 15 20 25 30 35
POROSITY[CoRS,%]
FIGURE6
1000
100
10
1
0.1
0.01
0.001
0.0001
0.00001
WELL B - COMPARISONOF RESULTSBETWEEN
THE FIVEGROUP MODELANDTHE
PRESENTLYUSEDMODEL
80.00 85.00 90.00 9s.00
.-.. --
lW.W i05.00 liMO
ii~.g) ~mnn
a2u.ul .l
DEPIT-I[mares] - only last digits of depth figures are shown
WELL D - COMPARISONOF RESULTSBETWEENTHE FIVE GROUPMODELAND
FtGURE 7 THE PRESENTLYUSEDMODEL
DEPITI [met!es] - only last digits of depth figures are shown

Vous aimerez peut-être aussi