I’%-ilmaiasoclstvofPstfdwnEfmhou9

SPE 30153

DRILLING FLUID PROGRAMMED CAN REDUCEMINIMISE HOLE PROBLEMS : CASE HISTORIES

AJ T~ PM Collins, MD Jacksm p Fo-

caPwwlsss.sOdwYof~~. I*km-~ti~tiatiS=W~WWhWW_W,SR Vimmn.1-3_lW

HOLE EROSION ANDOFFSHORE VIETNAM

This paper describes technical studies undertakenatBP ExplorationTechnologyProvision&n@ and theapplicationof these studiesalongwith tmnsferof bestpractice and local knowledge to solve two spec~~cproblemsencounteredin drilling offshoreVietnsm.

1.

2.

Hole instabilityin MOW dastics : hok probkmsMated to drilling shallowClaystonesectionswithwater based muds were behg encountered;tighthole, bit balling,stabiliserballing,overgaugehokThis was causinglost time and increasedcosts. Atechnical audit involving the BPX TechnologyProvision Centre, Dowell IDF and BP Vietnamalongwith technologytransfer fmm the NorthSea@CXStiOIIS ~OWed the dri~ing fluid used on @problem section to be modified/optimisedresultingin much improvedMlling efficiency.

Hole erosion in deeper elastics : the hole erosionencounteredin dcepir clasticsectionsof wellscanlead to several problems, namely the failedattempts to secure adequate wellbore evaluationinformation in the reservoir section. Cores wemavailablefrom problemformationsand tkse wexcused in geomechanicaland chemical analysis to~rtak ●h* mmm,m.44 ●- msW9M.- m..-#l -— use w- VA u& puuscws. &zaaGuw thefindings it was possibleto make recommendationsto the drilling fluid programmed and drilliigpractices, which have lead to enhanced drilliigeffkkncy and much impmvedwellboms.

BP Explorationcommenceddrilling offishmeViifnamin Decemberl~operatingin BX 526andlllLSeveral technical challengeswem experiencedin thedrilling of these wells. In an attempt to optimise onMUingperformance,severalstudies wem undertaken.This paperdetails two such studiesundertakenby theBP TechnologyProvisionGntre in Sunbury,UK.~first study was to carry out a technical audit ondrilliig fluids being used in an attem t to improve

t!’drilliig ptWfOIIMIUXill Sh~OW ChS C fOIllMtiOIIS.The SeCOIAdstudy Concerned~ ObSSIWdproblemofwashed out holes in the deeper clastic formationswhich had resulted in problems with adequateformation evaluation. This paper describes theproblemsas idendfie&the methodsused to determinethe cause of the problems and discusses theconclusions and recommendations resulting inimprovedperfoxrnance.

Oneofthe moxcpnMematici@esectkms OnwellsinVietnam(excludingthe loss control problems in the..L-.+-1 :- ~. ** fis- I mm . ..ti -- .L—.. _L A.bal Wuuuza ] La w at IL I-AIA SUwun Ulruugn cue

shallower elastics. These sections w~ orkinallydrilled using a gel/CMCsystem which was graduallyupgradedto a KCL/Polymer(PHPA)system.

Problems encountered on this section were due to“water base mud” beiig used to drill shallow claystones. (Bit balling, tight trips, reaming, shskerblinding, stabilisers caked up). This was leading toexpensiveperformance

mud bills, compromised drilhgand poor quaiity hoie.

References and figures at end of paper

2 DRLIJNG FLUID PROGRAMMED CAN REDUCE HOLE EROSION SPE 301S3AND MINIMISE HOLE PROBLEMS : CASE HISTORIES OFFSHORE VIETNAM

Variousanalysesof the problemhad b caxriedOUbbut no definitive answer and fonwrd programmeagteed.

As such an independentaudit was commissionedbyBP Vktnam to audit the drilling fluids used on thisSectionintbepast, fully evaluatethc problcmin co-operationwiththe mudcom~ytiauptiaset of recommendations for an optimised mudprogramme.

m audit includedCompmhensivedata collectionandanalysis, intaviewing appropriate operational staff(including service company personnel), cutingsampleaualysisand inhibitmn&at&

The use of the more inhibkive mud systems didSmUDSF*~ u?vz imprnvd clfilli~fz @ff~~e*~ (SCC=wrw- e“ - — ----- -- —-0Table 1):

<i) the average rmmsmy vaiues “Sere gmwra!iydouble those values seen with gel/CMCsystems(see examplevalues below).

lam

(1.TOBS)

i%%.

(LT OBS)Kan&ylW

sxlsr#Lmm&sl

6.90 23.s

lwfl 3a4

7.15 W.31ss 84.3

7.99 623726 172

S.33

427

3.s4263

4.9s

w

&16357

Incidentsof bit balling were more frequentwhenusingthe Gcl/CMcfluids.

The inhibitive muds used were partiqdarlybeneficial in the mudstones above carbonatefrom a geological stand point. The cuttingsobsexvedwhen using such muds wem able to beused for geological interpretation- an importantfactor when looking for top carbonate.

The stilt report concludedthat

1.

2.

The major formation problem was related todispersionratherthan “swelling”.

The inhibitedmud system based on PHPAshouldcontinue to ‘be used ivitt qXimkd pohy=ermakeup.

3. The alternative encapsulator system (hem CaIlcdpolymerPA-X)pm ad by the mud Comwy as

8anakemativetoP A looked~.b~Bphadnofield datatosubstdatci tsuseblwowsections.

PA-Xis a shorterchain PHPAnormallyused in hightempemture mud formulations. Its usc in this lowtemperature, shallower depth a~ication wasrecommendedon the basis it is earner to add andmaintain concentmtion levels in high clay contentmud systems. The viscosity humps associated withaddingxegularPHPAaxeconsWmMyreduced

BP had no field data on PA-X used in shallowsectiol&butstatoildW ‘Illeyhad evahlatedirauscinStatfjord and concluded that its performance inShsIlowsedimentsjustifiedits Contintmlu=

Withthis fielddata available,a trial of rhe KCL/PA-XSp wasqpmved fix the OS-2-Bac-lXwelL it wassuccmsfulandhasnow beenusedon WLT-lXR and06-LT-2YL‘IIMtabulatedresults am indicativeof thesuccess -themal dataisthcpoaitivc fdbackfhmthe rig. No Wng shaker screens. minimal bitball@g, no reamin$, gauge hole, no wiper triprqumwl after Iogg-mgand tie prob~efis rwmizgcasing. ‘l%cfirs twipertri pisatilltig ht-butthoseofyou familiarwith the old days using waterbasedmudin the North Sea will rememberthat! The problemshavenot disappeared-they arenowbeiimanaged.Themsldtis tbatwcammwa bktodrillncarl twice

{as fast and deliver a better gauge hole or theexplores to log - a definitecaseof win-win-we pay asmallcostinhighermud bills.

h summarythis auditillustmtesthe following

1.

2.

The process to achieve an improvement inperformance involving a Technical AudiL MudContractor involvement and the use of theAllianceresource.

The success of the KCL/PA-X SyStCm in thiS

BP Exploration’sTechnologyProvision in Sunbury-on-Thames, U.K. were asked to evaluate the poorwellboreperformanceof Well 118/BT-lXin Vietnam.Thewellprovcd to bedmcultto log, andobtain RFrsamples from. due to the frequency and size ofboreholeenlargementswithinthe reservoirintcrvaL Aprogram was undertakento review the existing dataand perfom some basic labomtory tests on cmspecimens in order to better understand the combehaviour.

The objective was to detemiine the cause of theborehole crdargcments in the reservoir zone which

SPE 30153 AI lTVYNAM, P OCXLINS, MD JACKSON, P FORMAN 3

presented diffkultics for subsequent geophysicallogging and RFI’s, and recommend preventativewasumstobetakenforan ewwellinthesamea

xistoIical driMngdat& inthcfmlno fdlilling P.Sndlogs, werccxamincd togainanmkmmdmgthe events that occurred regarding this WC8:@@@~ lop ~*, pdcuhrly * calliperiogsand gamma mytracts, to determine whefe theborehole enlargements occurred and what their- ~ we=.

Comleplxtsandgeologicalmportswcmstudiedto tryto Underaandthe rock’s‘kha-virmrin ressnse ●& ●*_O~-bo*rwh thcsiscamodeof

Ffdhue for the wellbme. Oncea ew pssible scermioswcm developed, laboratory tests on specimens ofsandstone core were performed to determineadditional physical properties for * tmublc-pronereek. Thcresldtingd atawereu scdinsubsquentanalysesto identifythe probablecauseof the wellboreinstability.

A brief summary of the well geology is providedin* followingtabk

Twocoleswere lecovcledfmmthewelk

1 3150.00-3158.78 m2 3547.25-3565.25 m

One hundred and fifty sidewallcores were also takenwithin the resewoir (1980-3792M). They confirmedthe generalreservoirdescriptionof beii an immaturesandstoneto silty sandstone,withmudstoncstringem.The sandstonecom has low penncabilitjt 5mDfor theupper core, and 0.5mD for the lower core. PorosityWithintherock isnotall interconnected.

Formation Pressure and Mud WeightSummary

With the formation above the reservoir beingoverprcssured, the reservoir interval (31OO-38OOM)was drilled with a mud weight of 1.32 S.C?.(withnormal foxrnationpressures) which msultcd in anoverbsknce of 15(K)to 1836 psi. It is twmh notingthat WdbOmiIBtab@ PrObbUl wouldIIOtJIOIM~ybeexpected withsuchahigh wubalana.

Thcrcpcat connation tcatcr(RFT) wassuccasM inobtainingpressms about 50% of the time. Thh highfhilumratcwa sductoscalfailur cwhichisattdhabletothcirregular borehole shape after mhgemutMud weights for the interval tested (3107 - 3787m)wcm a codswnt 1.32S.G., awl oquivaktt finmathfluid densities were reported as 0.98 S.G. SOMChigher Vabs of EFFD at the topliill#V&uJhave bcendue tosorncsmallovpoor sealin&

The formation was generally described as tigh~despite the fact that many =S we= withinsandstones. The exception was the sands- at3350Mthatwasdescribedas moderate,and producedfiuid Sarnpies.

X-RayDMhction Analysis

Thi5‘XR *-t V@E (iOiE ‘~ tbtti Sil W8 of “hmineralogicaland structuralmakeupof the ~.

—

I Mineral 1 Weizht % 1

Iouenzlsll

lDokxnii127[IhficallollFeidslmlo91

Cation ExchangeCapacity

Thecationexchange capacity(CEC!)isamcasumofthe ion-exchangereactivitybetweenan adsorbedsolidandasolutiom The CEC!isuscdmcasumdinunitsofmini-equivalentsof ion per l(K)gof solid(“meq/lCK)g”).Typicalvaluesof CECam listedbelow

The well is normally pressured in the Bien Dong%rnation, and at the top of the Quang Ngai.orrnation (1238m) the pressure slowly climbs,reachinga peak of 1.35 to 1.43S.G. at approximately2800M. This falls to ~rmmd/neneurcs of 0.98 S.G.by 31OOM,ad remains

4 DRILLING FLUID PROGRAMMED CAN REDUCE HOLE EROSION SPE 30153

AND MINIMISE HOLE PROBLEMS : CASE HISTORIES OFFSHORE VIEI?V$M

I Clay Mineral I CEC (meq/100g) I

Monimailkniite 100

mite lotoa

I Kaiiite I 10 IThevalue of CECobtainedfiOmti~ coresample (3551m) was 2 meq/100g MBT (methyleneblue test). While this is extremely low, it does notindicatcthatthe rockisnon-reacdve -ti-krefkcts thereactivity ofthc--ple andlmtofthe clay portion of the sample. Although thissandstone is mechanically atmng, it appears to befiiablewhich wouldindicatcthat thcsandst=*itsatren@ fromgrain tograin coritaCt8ridnot fromStrongCementation ‘Ihe grainsconsistof bmt quartzand dolomite, therefore the reactive elements amconcentratedin the bii cementationof - rock‘fldscmldmeanthatveryr eactive~y~-inticementationam susceptible to chemical dexoaresulting in a breakdown of the cementation and a~oftimti~mx.

In additionto the originalcore studies,newtests wemperfoimed for this study. A dozen rock strengthtest~e~ ~rffiim-~ nn cn~. p~ggs Ln Qrder go own a

●.“--- w.. ---base-he rock strength (unconfined compressivestrength = UCS). Half the samples wem steeped indriuingfdtratc atwellmpe~d-mmtithe Ucs had changed. Mud ftitratc tests wereperformedon cm disks. The results of W -and their inteqxetatiomare Summarisedbelow.

An interpretationof the mud filtrate teats, relatingthedevelopmentof the mudcaketo the mquimdUCS forthe m~k, is pit-i-d bdoii h thi$OdCid ‘=rmS,fOUQWd by an examinationof the laboratorydata.

One-DimensionalFiltrate Flow

The approach taken in examining the effect of adevelopingmudcakewas to considerthe transientflowfrom a wellboreas a pmbleinof impededsinglephaseflow, where the flow impedance was dynamicallyincreased due to the accumulating thickness ofmudcakeon the boreholewall.

In a onedmensioiial analysis, the referencestate forfluid flow from the drilling mud through thefoiniation is the case of fdtrate with no solids.Assuming that the filtrate has displaced the mobilepom fluids and has reached a constant saturadon,theflow rate across a given length should be constant.

This assumes that them is no cordiltrate interaction ‘that wouldaffectthe fluidflOWchamcteriStiCS.

The Darcy’sLawflowve@ity woiddbe-m

v = -k(Vp- yvz) [1.1]

whim!Vp isthepiessm gradienty isrheweight denaitygradkntVz istb elevation gradient

k isthemobility

The mobilityis definedaiK

whimk isthe absolutepemieabilitykr isthcmladve penneabilkytofilh’Wep istbfiltratcviscosity

If gravity effects are excludedas beii second-ordereffects,the flowrate canbe defimd as:

q=vA [1.3]= -k Vp A

= -k ApAMA

q = T Ap [1.4]

whereA istkcross-sectiondamaT isthetmnsmhm“bilityof tk plug to filtmteAp isthepmssum differenceacmsstheplugAl isthelength of theplug

Equation1.4 states thaL for a constantpitmure dropalongthe plug,b flowrate will be _lled by W

-—:..’:l:k:k. A ●k ..-.1. ●- fjj~,g& flow.&fecdve traaISIJIISSiUAAILYUf-= -P w

&neDhnensionalFlowwith Accretionof Mud Filter

The addition of drilling mud solids to the filtratemsultsinan impedanceto filtra&flow astheaolidsarcdepositedon tk face of the plug. These solids aredesigned to plug ttw PO= throats, thus reducing theeffectivesuesfor fluid flowinto b cm. Iii addition,the continuous accumulation of additional solidsforms a mud filter cake that has two effects on fluidflow fi~ the additionalmaterialresultsin a toxtiiousflow path for ffltratc passing fi’ointhe drillii mudinto the core plug, and secondly, the resultingincreasedpressuredropacmssthemud cakemsultsincompactionof the mudcake,whichfurtherreducestheeffective pe!meabiity of the mudcake. A schematicof the flow rates is shown in F@e 1, in WhichtiE

SPE 30153 Al lWYWW, P COLLINS, MD JACKSON, P FORMAN 5

_fil-tiW~Mc$mtimfirbcaw of pure filtrate flow and fdtrate flow with anwxeting mudcake.

-

l\/--

Teff = q(t)/AP [1.6]

Sithemudcake andplug amacting inscries, theeffectivetmnslnissibilityis Complisedof the twosepamtetransmissibiities:

‘Teff)-1 = (Tmudcake )_l + ~cofeplug )_l [l~T]

Ap/ q(t) = Aprnudcake/ q(t) + Apcorcplug/ q(t) [1.81

By plotting Ap / q over tbe @s&a curve similar toFigure 2 should be obtained. The y-axis interceptrepresents the condition of zero mudcake tMckncs&which ahouldbe thesame as the caseofpurefiltmteflowwithm solids.

m............................The exponential form of the cume is due to theaccretionof mud solids on the face of the com plug asffltrate is forced throughthe sample. If it is assumedthat the thickness of the mudcake is directlypmpotional to the volume of fdtrate that has flowedthroughthe plug, the mudcaketransmissibilitycan bexpressed as a ilnction of fdtratc flow. From

Atylations1.3 ml 1.4:

T(t)mudcake = u /lU(t) [1.9]

= u /(gQ(t)) [l.lO]where

Qistbecumldath?efntrateflow,ma=mofmgis a pmpordonalltyconstant

By combin@ Equations1.10, 12 ad 1.8,

k = g Q(O ‘NOmudcake/A [1.11]

Notcthatthe equatkmassumes thatthe~inhmaincoribu%.HOWOVCG a8t&volumeofmudsolids deposited as the filtercake increaaea, thepreamre dmp acroas the mudcake increasca, whichahouldcompactbmudcake anddecma8eiU~permeability,as seen in Figuxc3.

MOBILITY m MUDCAKE ACCRETION

TIME

Theory

Sithetotal ilowthmugh thesampleis thesameforany cross-section. the following relations can be-.

q(t) = Ten A~@ = Tmud~ A~udc~e= [1.13]Tmreplug Apcorcplug

Mudcakeeffkkmcy “q” is definedas tlw effectivenessof the mudcake in maintainingtk total overbahmcepressuredmp withinthe mudcake:

q= @mudcake/ APtoud= T~/ Tmudc~ [1.14]

From [1.7’J,

q= 1- Teff/ Tcoreplug [1.15]

\

6 DRILLING PLUID PROGRAMMED W REDUOE HOLE ER=ION SPE 30153AND MINIMtSE HOLE PROSLEMS : CASE HISTORIES OFFSHORE ViiFiliAM

-.——

The transmissibilityof the coreplug can be obtainedby extrapolating the total transmissibility to timo=(lwhcnthemudcake didnot~m Acontinuousplotof

can be obtained by -Wing *Coreplug Sndti total tnmsmsibility.

There ●re some assumptions made with theseequations. First, it is assumed that the pressure~mtim ~ * -x h - ~y~Y@:thatthe pmssumtransialts~thc1800psl Apto~acrosst hesa@ebave occurredinstantaneously. This simplification wouldoverdmatc the complug transmissibilitybecause itWouldinitially beeasier for filtrate to flow into the‘tipiee-d =~p%tg. S~nd. it SSSSUMU ‘M ‘W

value lernains constant throughout* test and doesnot change due to dissolutionof soluble minds orby the formation of an intcmsi mud- –Wittfinesdistuw by, or transsrtt by, the fiitre

‘I& efficiencyof tk mudcakeaffects the stabiity ofthe wellbore because it is the effect of the stressapplied to the sandface that strengthens a frictionalrock such as this sandstone. When the mudcake isineffective,tire is no appliedconfiningpressuxe,andtk sttrengtt ef ttbe rock is greetiy reduced. Thebenefit of the confining pressure is of greaterimportancethan the intrinsic unconfinedcompmasive- -~ n rne~ -r - -n-l. .:--- ●L. d-a h ae4:sW3mi%lgul [ Uu) UA & lU&A Sulwz USG isuag%a ~cubaa

with a confining presauxcfar exceedtk UCS.

This can be seen in F@e 4, in which a Mohr-Coulombfaihuc criterionis showw

~g= 4

MOHR-COULOiiiii9 FAiLWtE CR!TERiON

SHEARSTRESS

{ /

Ucs u,‘3 NORMALSTRESS

confiningstress -

N.B.: arennuedkdk~

The valuefor the Mction@e, $.- in the ~Y@was only an estimate since its determination wouldhavemquheda set of triaxisl tests. However.h m~range of friction mgicsk32m 40degreesf0rw@to strong sandstones. The Ucs values Obtakd fromthe labomtory testing on 1 inch diameter plugsindkate that the sandstone is moderately strong, asseen in Table 2. Note that the UCS values for all

samples(unmated, -, -treated) ~ -~Ythessme. Whencxammmgthe mm, however,it wasftiable, indicatingthattherod ccouldbemscept ibletodamage,such as erosiom

~

DEPTH Ucs Ucs iuFtratt 1(psi)

3151.7

3153.0 6430

3155.0 6s30

ww n.“. .- ~~~ 65609CC19aaaa.t ~

3555.02

1 TJIeBesmlJplalVcmtludKdwitb5pllra vobmnuoffiltrnaaIlsow, tlKah8kl uIbAttcmpEmm Witt Ismfbwfalhur.Mt6amlin&ucs teuslveleda18.

2 ~ebaa-~

. ..A.* ●h. * Anm ammlamd T“TG,*4 ~~ag~By ImuwAu& w Aaswuwsmcu4&u —

of a rock can be determined for any confiningpressurefrom the equatiom

al = q (l+si@y(l-si@) + Ucs [1.16]

Convemcly,the UCSfor faihue can be dewnined foranyplincipslatlessstaE

Ucs = 61- cJ3(l+sin$y(l-sin$) [1.17]

In situ stress state

Thestmssstate attheboreholc wauwasdetemincdbyassuming an overburden gradient of 1 psi/ft,normally-pmsmuedformationfluids for the nxstxvoirelevatiomanda Poisson’sratio for sandof v = 0,2S.

QuiescentDepositionalBasin

The common assumption for the stress state for abasin is that it is a result of the accumulation ofsediments.andthat thehorizont dstmssesamarcsultof the clastic response of * mck to the overlyingweight. Tk Lsn Tay IWWIVOiris not in an extensionalor compressionalfield, thercfom this assumption isjustifkt. Also, in @ absenceof any indicationthatthe horizontal stresses arc very dlffercnt, theassumptionbas ‘beenmadethat they arc

Tai. ThiSis

rarely the case, however if the stress “fferenceissmall, the error introduced will be minor. In thisMtance,thcinskusuesscsarw

SPE 30153 AJ TWYNAM, P COLLINS, MD JACKSON, P FORMAN 7

%= Yobz [1.18]

f = Yeffdz [1.19]

m= W(l-v) (t%J.pf)+ pf [1.20]

Wix!lt

* ist&totalvertical atress

Yob is the overburdengradient(e.g.: 1 psi/ft)isthe verticaldepth

if is the ftmnationfluid pmssumYeffd is ~ @v~~ fom@on fl~d ~1~a istitotal horizomalatressv is Poisson’sratio

=:WM, ~h~ ~~’f’ ~~ t~As ~ ~~~~ ~gg ●Ms ~s~ntt nn

mudwtme at a depth of 1978m was analysed,assuming an overpressum of 1.29 S.G. With therecorded LOT gradient of 1.74 S.G., and usingEquation1.20,the Poii’s ratio was foundto he n =0.30 whichis reasonablefor a claystmw

31e drilling of the Ixmho]e alters the stresses in therock fmm the in situ mess state to one of stressconcentrations around the borehole. The verticalsucss remains essentiallythe same, but the tangentialaessesarc raisedwhiletheradialstmsses welowered.

rIowever, the application of a radial supportingpressure through the use of a heavy mud weighrandeffective mudcake reduces the stress differencebetweenthe tangentialand radial stress. wltbut suchsupport, the borehole usually fails in a classicbreakou~ with failure occurring by a constrictingshearingof the mck mass in towaxdsthe Ixxehole.

In additionto the changesdue to stress concentmtionsare pomelasticeffects. ‘Illesearc @ changesinm~Stressasamukofchanges inthe fluidpressurerock ASthese fluid p=- VW, t.k pIOpO1’tiOllOfthe total stress borneby the rockmasswuies inversely,which causes stress changes or displacementsin allthree principal directions. With an imperfectmudcake, the fluid pressure changes resulting fmmthe influx of fdtrate alter the rock stresses at thesandface. Since this is wherestress mncentmtionsarecommonly the highest, the state of stress at thesandface with an imperfect mudcake will be~~~~p.~.

The three principal stresses at the sandface are(McLeaIu1988):

62 = C$z+ (1-q) (pw~ - pf) (1-2v)/(1-v) [1.21]

Ug = 2* - pf+ q -. pf) (1-2v)/(1-v) [1.22]

~r = ~~ [i.23j

The pore fluidp~ justinside thesandfaceis

Psf = pf + (1-q) -. pf) [1.24]

Themfom,the three principal @ectfve stressesat thesandface,meaning that portion of tk stressm bornebytherockmatd~as opposedtothefluidp~am:

CJz’ = CJz. psf [1.2s]

00” = q . psf [1.26]

The principal effective stresses c1 and cr2 am tkmaximumandminimutn of these. T& UCSatf&ilurecan then be calculated using Equation 1.17 for anylevel of mudcakeefficiency. Note that this type ofanalysis assumes that the sandstone behaves in alinearly-elasticm-, and that the pmssute transienteffectsattistatofthe testamignomd.

Such an analysis was done for the fivemudcakejfiltratetest samples. An analysiswasdorMofthe varying flow rates, in order to determine thetransmissibilities. These were used to cahlate themudcakeeffectivenessover time, which were used tog~d~wxxe prcssum and effective atmsses at the

Whh these, the UCS at failure wasCalculati.

‘fhedrilling ilhratelosstestswexeconducted on fivedisks of core. Three wem obtained fmm the uppercored intavsl (3151m,3153m,3155m)and two fromthe tighter sandstone at the top of the lower coxe(3151m, 3151Ji7m). Most of the lower com wassiltstone4mudstom,with the mompermeablesdstoneappearingat the top of the intenfal.

Drillingmud of the type used in Well 118/BT-lXwasused for these tests. Core disks of 25mm in heightwemprepmd for the flowtests. Tb temperahuewasraised to 124°C (as compared with the estimatedbottomholetemperatureof 15W’C)and a differentialpressure of 1800 psi was applied. Flow rates werecontinuallymonkorcd.

1 Upper Core (Zone A)

The specimenfrom 3151m is anomalous in that theobsemed flow rates me four times higher than theother two specimens. ‘IIdsmay be due to somelocalintemonnech“vityof pore spacethat is not extensiveina larger sample,and is themfomnot mpmentative ofthe Upper come c&tairdy * mo mmuc~ htransmiasibilky during the test are indkative of aninczeasein tk permeabilityof the specimen/mudcake,

8 DRILIJNG FLUID PROGRAMMED CAN REDUCE HOLE EROSION SPE 30153AND MINIMISE HOLE PROSLEMS: CASE i-WTfXtiES GFFSHGRE ViETWW

whichwouldmake the teat resultsfor this sampleverysuspect

llie specimensat 3153mand 3155mam mom typical,in that the transmissibtity demases at a slower ratewith time. Mudcake efficiency is at 80% within 2minutes , with the required UCS dropping from aninitial5417

l??to an averageof 1285psi. Recallthat

the average CS for the upper core was 5748 psi (sixsampk) as seen in Table2 above.

A SUJb~ty@ySiS Of thiS h~d WMpCtiOllllCd,assuming mudcake efficiencies of O%, 50%, and100%. ‘IlmrequiredUCSVdUCSwere 5403psi,2798P& and 193psi. Clearlythe confiningpressureOfthemudcake adds considerably to the strength of therock,

2 Lower core (Zone B)

~-—” ~Thh twe spewuws 4.. 1..-. - .—m. ●- ha”-Uulu IrIE Suwfa wrG appGas w -Vw

evenhigherpenneabilitiesthan the uppercm, as seenin tk volumeof fluid loss qxnted over the test after20 minutes. This is contrary to the previous andcurrent physical laboratory core testing that found afactor of 10 difference in permeabiities. with theupper cme beii more permeable.

Duringthe test,both samplesbehavedSimiiy in thattiMdmp in transmissibiity with time was muchmore~auti. Mudcake effkieiieies develeped incxesiow~ye reaching 80% after 4 and 6 minutes. Thereqmred UCS strength at O%effmiencyis 6098 psi,and for 80% efficiencyit is 1403psi. lhese required= ~.~r - for the uppercme becauseOf

mcresseln well depth The Predictedstress levelwith no mudcake is higher than the average UCSobtainedftom the iower core was 5572psi (cxciusiveof the mudstrm results at 3555m). Required UCSvalues for O%,50%, and 100% mudcake efficiencywere6089psi, 3153psi, and 218 psi.

ONSOND~

The precdng sections of this paper detail the teststied out on tlE core samplestaken from well 118-BT-lX. The major points to consider from thisanalysisale as follow.

*

*

The Drilling CompletionReport and the calliperlogs were used to identify the severity of thewashedout sandstonesections. A typicalzone W=

identified which was relatively stable @me A :3151m - 3155m) and another typical zone which&55~ rienccd washouts (Zone B : 3551m -

Y.

Samples were taken as required from coresavailable at Sunbury from these depths to carryout severaltests.

* XRD analysis was carried out on the samples.These resukscanbeseeni nAppendixA. Oneobvious conclusion is ~ the samples from thewashedout section ~ a ~@ pememageOfdolomite (-10% up to 27%) when compamd tothe in-gauge section (~%). This was an~ptiti-yconcldons-hwnthat the highly soluble dolomite matrix in thesandstou couklhavebeen leachedout byddllingrnudf~, hence wdmning the rock strenghUnfortunately,this WU not -y CCMfifMedbylater unconfinedcompressivestrength(UCS)teats(see below), but we believe the high presenceofdolomite in the sand matrix could have possiblylead to an increase in filtration rates, andsubsequent reduced rock strength as the sandmahixwasdissolved

-l@mItswereas Wi-m

‘= -ZIMA: Iowa SMspci 5mD3131- 31ssolkB: Washehul Ss96pd M@3351-333%

The averageUCSresults fkomboth samplesindicateda very strong sandstom. An expert opinion suggeststhat any UCS value above 2000 psi indicates a

Alt&J@ “. —mnewt-mt ~&m~* _waUpMUSb th. esnrk$$ in 7= R-——.

hadlowerucs, thedifference wasnotlarge enoughtoexplain the erosion which had taken place.

* Using the same cwe samples, & effect of mudfiltrate onthecmsw asinvestigated.A mud wasmixed to simulate that used to dlillllthe original...-11 .-4 el+,ne- -fill-m*-#l Amm ●WGU mm AuuaMz WUGWW *awMm ~*& p~~

filtrate was then assed through the cores (1”rdiameter, 2“ lon and UC!Stests carried out

hbefore snd after trate invasioxLThe FracTechreport Conchldd that thm was no evidence Qfmud filtrate interactb with h samistom_either in the core sample from the in gaugesection (zone A) or samples taken fkom thewashedout section(zoneB).

* Further com discs wem cut (25mm) fmm bothzones A and B. Dynamic fdtration tests werecarriedout on the HuxleyBertramwith simulateddriUingmudasusedinthe oxi$nalwelLFmmtheresults obtained,a geomecbamcsanalysis showedthat sand cores from the upper in gauge zone Aachieved80%cake effectivenesswithin2 minutes,indicatingthat the cake buildingefficiencylead toearly strengthenhg of the mck However,zone Bshowed a much slower rate of cake buildingefficiency at identical parameters (124’C with1800 psi overbalance), due to its reducedpermeability.It WSSCOnchlddthat & washedoutzone B cores showed poor mud cake buildiigefficiency even at such high overbalance. This

SPE 301S3 AJ 7WYNAM, P COLLINS, MD JACKSON, P FORMAN 9

would have lead to a reduced cxmfhdngpressureofmud-to addtothe rockstmngth

seems probable from the above results that any&ture attempt to minimise hole erosion needs toConcemmteontwo key issues:

- optimise mud type to minimise chemicalintcnwtionwithh Sandmatrix

- optimise mud formulation to enhance filtrationcontrol and attempt to (a) reduce filtmtion rateseven further and(b]assist intbe cake bdlding-V of the mud system.

other relevant studies

A recent study by P. Page and M. Hogan at X’fTSunbury(ref.) looked at fluid loss mechankmsunderdynamic conditions. This report describes workcmicd out on ~M’s (prin@ly KCl/ polymerm-)to establishthe impactand xelativesignificanceof thefollowingpammeterson dynamicfiltrationrate:

. mud weight“ mud solidsparticlesizec differentialpresaum● temperaturec mud rheology● formationpemmbiity and pmosity

The relevantconclusionsof the reportwereas follows

1.

2.

3.

4.

5.

7he dynamicfdtmtionrate of Ka / polymerwaterbased mud was strongly dependent on thetemperature, mud solids wntew mud rheology,ahearmte atthefilter cakesurface, and the porethreat size distributionof the formati=

By Comparisondynamicfiltmtion was insensitiveto changes in differentialprcssum over the range2S0 -1,800 psi. This effect is atuibuted to filtercake compression.

Defining rock samples in tcnns of permeabilitycan be misleading in

aredicting filtmtion

behaviour. Comparing uid losses throughsyntheticand outcropcores has demonstratedthatthe pore size distribution of the mud is moreimportantin influencingthe filtrationrate.

Fluid loss can be significantly reduced by theformationof internal faltercakm. The Ulty of amud to form internal cakes depends on the poresize to mud solids panicle size I@o (PPR).

The results show the importance of determiningthe pme throat aim d~tribution of the formationsbeing drilled plus the particle size d~tribution ofthe mud solids if an assessmentis to be made of~ ~lff=uva= of ~o~o~ ffitoa p&information can be used to identify whether anintexnal ftir cake is likely to form, and morepmactively, whether an adjustment to the sizedishibutionof the mud solids woulditnpmvefilter

loss control. AhhOUgh tk Abrams (ref. 6) ruk-of-thumb is widely used as a guideline tominimise solids invasion, the test work resultssuggest that to promote the formation of anintexnalfilter cake the size distributionof the mudsolids SIMXIMsimply be broad enough to overlapand minor the pm throat size distributionof Wformation.

6. ‘fbesolids contentof themudwas foundto betheparameterwithmost infhmce on the effectivenessofsomeglycols asafluid lossadditive. AtamudSG of 1.1 (23% vol. total solids), DCP 101 had

—- .

m affect on fluid loss, but at mud S0s > @ L2(5.4% VOLtotal solids), the addition of 3- 6%DCP 101glycolto a KCI/ polymerm~ -Yreduced the dynamic filtration rate by 10- 20%compared to that obtained with the base mud.‘1’lEaddition of 3% Dc!P208 glycol reducedtherate byoxdy3%. Similartests withstarchrcducedthe dynamic filtration rate more than with D@101 (a 2S% reduction) and this was achieved atmuchlowerdosages.

7. The re~lts suggest that two filtration controlechmsms may beinoperation whenglycols me

Lintimud

(1) The viscosity of the glycol solution is higherthan that of a plain aqueoussalt soluti~ hcace inthepoms ofthcfilter caketkfikrateinvaaionmteis slowed.

(2) When glycols like DCP 101 cloud OULthemicro emulsion formed may plug pores andcracks intheformation in the same wayastbemulsionphasein an invert OBMdoes.

8. The results have demonstratedti impoxtanccoftempemture in determining the rate of fluidinvasion. Until recently, reducing mudtemperahm has never been considered as apracticaloptionfor reducingfluid loss. However,Elf and Alfa-Laval (ref. 7) have recentlydemonstrated that cooling the mud at surfaceusing heat exchanges can provide significantbenefits in terms of improvedwellboreStabiity /reduced fluids invasion to be justified oneconomicgroundsfor certainweUs.

Forward Programme

Based on the above informati~ and a consideradonof current “best practice” guidelines, a fewrecommendations can be made in an attempt tominimii erosionon futurewcik

* Superior fluid loss controI is exhibited by oilbased muds (OBMs) over water baaed muds~lK.\ ~ t~~ $ m&mfiwmMk*t*Atn ●b fset\ VTuav4a)

that OBMs contain wat& ~m”@-e&”;Mz”~deform to produce a filter cake with very lowpermeability (micru darcy). Oil base muds alsominimii mudhock interactivity due to the inert

10 DRIIJJNG FLUID PROGRAMMED CAN REDUCE HOLE EROSION SPE 30153AND MINIMISE HOLE PROBLEMS : CASE HISTORIES OFFSHORE VJEfNJW

external phase oil. For these reasons, it isrecommendedthat oil muds should be the firstchoice drilling fluids for any formations withpotentialwashoutswhichmay be due to chemicalreactivity.

* Ho-wever,tise of oii ‘be muds may set bepossible or desirable due to economic, logisticsand/or environmentalconsiderations.In this case,one must quantify the potential Ml:lenllllfwashed out Sections within the Jextmcosts may reinvolved ifwashoutaoalinotbseradicated from a particular area,. fo~flammifipoor hole e:eanifig resdttrlg :+ rbv VU~.Wcirculating times and hole condoning poorcement jobs poor logging data or fommtionevaluation etc? Extra costs incurred by the-r ~ p-don for U* ~ -i ofoil basemuds may beoffsetinmany casesby theimpmveddrilling efficiency,reducedhole msionand enhancd formation evaluationperformance.This option should be seriously considered jfpossible.

* If oil based muds cannot be used, or are deemedto be not cost effective for the project, thenoptimizing the water based mud programmecanbe attempted. A number of options are discussedbelow.

* Recent advancesin ‘WBMchemistryhas seen theincreasing use of a new generation of WBMsystems containing glycols. Glycols am addedprimarily to promote shale inhibition, however,there is some evidence to suggest that they alsoreduce fluid loss. Fkld evidence also indicatesthat, generally, hole condition and stability isimpmvedwhenusingKC4GlycolsystemsoverMSinhMing water based muds. For this reasomKC1/Glycolhasbecomethestandardwater basedmud in XBU for technically challenging wellswhereoil base muds can not be used. Of relevanceto thiSStUdyis the observedtiUCtiOllOfiibti(mrates at high overbahce whenusing glycolmuds.

* Fdtration rate reductioncan also be achieved byother means. Improved filtration control can berealised by using various mud additives inconjunctionwith or instead of glycol. Starch is avery effective additive and should be considered.In Vietnam, however, the strong possibility ofbacterial activity may preclude dependence onstarch alone for effectivefiltrationcontrol.Use ofasphalts will also be effective hem, paxticadarlywith the BH’I”s to be encountered in theseformations. The downside of asphalts may beformation damage considerations.This could beinvestigated further using reservoir core ifrMxsmry.

* It is recommendedthat effective bridging agentsbe built into the mud system. One of theconclusions ftom the geomechanics analysis of

the core from 118-BT-lX,was that~a!lWprm&to washout had a much lower

xhence a reducedcapabiity to ectivelypromotea strengtheningwall cake. One opdon &re couldbe to eff-vely increasefiltrationrate of the mudSystem therebyresultingin an ~ efficiency*WMid a mud oaim The downsideof this wouldbe that fdtrate invasion would be increased,p#oyti;M&l& to an increase in chemical

thesandame. Abetter optioqwe believe, wouktbe touse asuitable bridgingmaterial to promote a rapid wallcake deposition,thereby effectively Smngth@ theweUboreoAmktureof bentoniteandfineOakiumCarbqtewouldbe recommended.Concernsover formammdamageif using this method would be over-rulediftheweUwas to beperforatecLasthe~wotddmost certdnly exceed the “damaged”xone. Thismethod would not be desirable, howev~, if bWellwasto becomplaedwithoutt herequirementto perforate.

* Use of encapsulating polymers, such as polyanionic cellulose (PAC) or poly-acrylamide(PHPA) would also aid in strengtbning of thewellbme by ~moting a polymer coating on -Weiibm, ana Secondiy,by Viacotifying-b mudfiltrateto effectivelyreducefiltration

* Anothermeans of minhising wellboreexosionis‘@@fitFOithe fkiid IileCiUdd SCtiVig~. M ‘4

thoughtby someexpertsthat wellboreerosioncanbe hdUCCd by tuhdent fk)W of tk Snnuhr fhdd,This can therefore be minimised by control ofannular hydraulics and design of fluid rheologyto promote laminar flow. A recommendationwhich can bemadeto achieve this would betoincrease thelow shear ratet&reby allowinga decreaseinflow rate without compromising effective holecleaning. There are several mud systems on themadmtwhichcan be appliedto this aituatiowforexample, a Mixed Metal Hydroxide (MMH)systemand a polymershearmudsystemsshowhighandhavebeenusedby BPXin severalapplicationsto enhancehole cleaningat low flowrates.

The majority of the overgauge intervals of thewellborewithinthe reservoirweredue to washouts. Inpardcular,the large washoutat a depth of 3560mwasa resultof a drUming washou~

Itistempting tostatethat thestnmgth analysisprovesthat the upper com was stronger than the requiredstrength, whereas the lower con was no~ and thatexplains the difference in performance. However,itshouldberemembered thatthe analysisis basedonalinearly elastic analysis that commonly under-

SPE 30153 AJ TVVYNAM,P COLLINS, MD JACKSON, P FORMAN 11

estimates rockstmngthsbya f-r of3t06ttithmfore the mess analysisis not conclusiveproof ofV~bO~ instability.

Althoughan elastic analysis typically underestimatesthe true strengthof the weiltmre,it is liiy that thisfactor should be low for this well for the followingreasom

●

●

●

the rock is uncmfhed because of the poormudcake,and will thereforenot behaveplasticallyas it appmches faihue (the UCSspecimensfailedbitterly)

the lack of a mudcake will prevent thedevelopmentof a stable plastic failed zone at theboreholethat acts to supportthe stronger,unfailedmck farther from tk wellbore

the type of breakout anticipated with this stressregime is conchoidal vertically, therefore therewould be no inherent arching of the failedmaterial if indeed borehole breakouts areOCCmTing,therefore no post-peak StfCllgthwouldbe available

Rockscan withstandsomeshomemn overloading,andit is likely that the major strengthdifferenceis a resultof the differentrates of mudcakebuild-up. & seen inthe plots of the required UCS versus time, thestrengthening effect of the confhing stress is fararger than the need for a high UCS value. Effectivemud design is the solution to strength problems infictional rocks such as sandstones.

Lastly, it should be rememberedthat the majority ofbomlmleenlargementswere washouts. It is likely thatthese could be reducedin !iequencyand diameterif amom effectivemudcakewemused.

To this end, it is recommendedthat severalchangestothe drilling fluid programmemay be attempted.Tkuse of oil based mud would cenainly minimise anywellboreerosiondue to chemicalinteractivitybetweenMilin$ fluid and the formation.However,it is realisdthat tlus may not be a cost effectiveanswer.Therefore,it is recommendedto enhance the performanceof thewaterbasedmud programmedto be used in an attempt.-.w.

a.

b.

c.

minimise filtration rates of the fluid by use ofmud additivessuch as glycois,specialistpolymersand pelhaps asphalts.

minimise mechanical erosion potential byincreasing low shear rate viscosities, therebyaUowingreductionof flow rates to attempta morelaminar annular flow profile, withoutcompromisinghole cleaning.

addition of bridging agents and srmdladditionofbentonitein an attempt to promotemore effectivecake buiMingpotentialacrossthe Ssndstom.

1.

2.

3.

McLean, Michael (1988) PhD thesis, ImperialCollege.

Abrams, A. Mud design to minimise rockimpairment due to particle invasion, JPT, 1977,pp. 586-592.

IvhNuY,V. and Guenott~. Practical advantagesof mud cooling systems for drilling SP41ADC25732, 1993.

Table 1Semi-QuantitativeClay Mineral XRDResults

118-BT-lX3i5i.5m

IlliteKdinite+(

Semi-Quantitative

K-Hdsplrcslcitewrite

Table 2Whole Rock XRD Resuka

118-BT-lX3151.9m

;:3%2%1%

z

semE-Qammative

miteKdhlite+cbhriteTotal

Semi-Quantitative

Mica+mitc

K-RldspIw=

%100%

Table 3C!ay Mineral XRD Results

118-BT-lX3551.37 -3551.39m

1%100%

Table 4Whole Rock XRD Results

iiil-B’i’-iX3S51.37-35Sl.39m

L:2%1%10%

I06-LD-lX

1o6-06-LT-lXHDB-lX

I0S2-NT-lRX

F2-BAC-lXLT-lXR

P-LT-2X

LENGTH

232

66110491221

339S72

642

RILLING

4

6119

13.5

2.8

DAY

58

11095136

339163

229

CGSTI

1724

9071049737

294612

436

MUDSYSTEM

GELKMC

GErJcMcKCl&GLYMERKCLPGLYMER

KCY.#A-XKCWA-X

KWPA-X

Includeswipertrip forlogging problemsUptowcllcontro lincidcntIncludesintcxlncdi8tclogsExcludeswwulm

r

WntiIncadjustedfor re-drillofsection18”gaugehole18”gaugcholc -includmVSPlog17.75”gauge hole