You are on page 1of 12

Society of Petroleum Engineers of AIME 6200 N. Central Expwy Dallas, Tex.

75206

Use

Rate Flow of Short Term Multiple To Predict Performance of Wel Is Having Turbulence

Tests

Lloyd G. Jones and E. M. Blount, Mobil Research and Development Corp., and O. !i., Glaze, Mobi1 Oil Corp., Members SPE-AIME

THIS PAPERISSUBJECT TOCORRECTION @Copyright 1976 American Institute of Mining, Metallurgical, and Petroleum Engineers, Ir?c. Thispaper wasprepared forthe 51st Annual Fall Technical Conferen~eand Exhibition of the Society of Petroleum Engineers of AIME, held in New Orleans, Oct. 3-6, 1976,Permissionto copyisrestricted to an abstract of not more than 300 words, Illustrations may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. l%blication elsewhere after publication in the JOURNAL OF PETROLEUM TECHNOLOGYor the SOCVETYOF PETROLEUMENGINEERSJOURNAL is usually arantedu~on request to the Editor of the appropriate journal, provided agreement to give proper credit is made. Discussion of this paper is invited

ABSTRACT A method is presex~~ed for predictingwell performanceand analyzingcompletionand fractureeffactiveness. The new procedureuses productiontests, rather than buildup tests, to determinewhether turbulenceis a major factor in reducing a wents flow capacity. If the pressvredrop from turbulentflow is greater then expected,the well probablyhas an inefficientcompletion, For a perforated completion,this may mean too few open perforations; for a fracturedwell, it may indicate that the fractureis too narrow. The analysis procedureis applicableto most gas wells and to oil wells with high flow rates. Field cases are shown for two gas wells cnd one high-rateoil well, INTRODUCTION

has been appliedto a hi~h-rateoil well. It includesand exten s the techniquereportedby !3if their n value is two In Phipps end Khalil, particular,the procedures}mld provide a powerful analyticaltool in areas where most wells have high productionpotential.
q

An er,alytical procedureis describedhere for predictingwell performanceand analyzing completioneffectiveness of wells that have a significant pressuredrop from turbulence.* It is more applicableto gas well completions but The analysisprocedureNermits determina*In t~s discussion, we use the term turbulence tion of turbulenceeffectson completion to describeboth turbulenceand ali other efficiencyirrespective of skin effect end rate-dependent devia ions from darcy flow such The requireddata are lsminsr (darcy)flow. 1 as inertfal effects. either (1) two or more stabilizedflow tests, or Referencessnd illustrations at end of paper .

The procedurecan be used in wells requiring send controlmeasuresand in hydraulically fracturedwells to determineif the cross+ sectionalarea open to f10., into the wellboreis sufficient, It also providesan indicationof perforationeffectiveness in normallyccmp!.eted wells because an abnormallyhigh turbulence coefficientindicatestoo few open perforations. Incidentalto determination of a turbulence coefficient, the procedureprovidesa laminsr flow coefficientthat includesskin effect. If permeabilitythicbess is hewn, an estimateof skin effect can Le msde from the lsminarflow coefficient. Also includedin the theory is en ex@snation of the effectsof partial completion or a chsnge in completiongeometry!on presswe buildupresultswhen the turbulencepressure &op is significant.

USE OF SHORT-TERMMULTIPLERATE FIOW TESTS IU~EDICT PliRF7)IWMANCR (MWET.TS HAVTNf? . .... .- TUREUTJ!TUCE . . ..--. -. -- --------- .. (2) two or more isochronalflow tests. Flow rates and bottom-holeflowingpressuresmust be known in either case. Transientpressure data are not needed and bottom-huleflowingpressures calculatedfrom surfacepressuresmay often be sufficient.

Smi! -. - Ala
w -<.

P@ = flowingwell pressure~psi q = oil flow rate, STB/D gas flow rate, Scf/D Q= t = laminarflow coefficient. for oil or gas wells (definedin AppendixA) D = turbulencecoefficientfor oil or ga wells (definedin AppendixA)

The previousmeans of determiningthe turbulencecoefficienthave ?equired some means From 13q.1 it is apparentthat for gas wells a of obtainingkh; usua21y a buildupt~st coupled \ - Pwf2 P with severalproductiontests or a series of vs Qg has a slope of D, and buildup tests.y~~ These were used to determine plotof s a total. skin effect, s?, which includedthe 2Q9 ~ turbulenceterm. The St valueswere t,~+m plot- ~ = lifll For oilwells$ a plot of Q~OQ ted vs flow rates. Actual ekin effec, and s! turbulencecoefficients were then calculated P Ws - Pwf from the interceptand slope. The method -vs q has a slope of ll,,and q describedhere avoids the necessityfor transient data from a buildupor drawdowntest and ~ = Iim 4 permits a direct plot of pressuredata vs rate. The plots apply to both linear q-o q, The added simplicityshouldmake this procedure much more useful in direct field applications. and radial flow. With these values alone, flow rates can be predictedfor these completionsat THEDRY other arbitrarilychosen flowingpressures. Steady-state radial and linear flow equations includin turbulentflow areavailable in !? for both oil and gas wells. the literature~b These equationsapply directlyto stabilized producingweZls; i.e., wells that have produced long enough so that the pressuretransienthas reached a long distanceinto the reservoir. The equ..ionsare readily adaptedto nonstabilized we..s by using isochronaltests; i.e., short flow tests of equal time duration,with each separateflow test followedby a shut-inperiod tclallow wellborepressureto rise nearly back to initialreservoirpressurebefore the next flow test is performed. The isochronaltesting procedureshouldprove particularly useful in determiningwhebher the fracturewidth is sufficient in fracturedwells. A seriesof very short tests may be sufficientwhere.otherwise a long-termbuildupwould be required. Z@M?M A simplediagnosticprocedurecsn be used to analyzecompletioneffectiveness of both g s 8 equation, and oil wells. Using Forchheimerts it is shown in AppendixA that flow rate and pressuredrop can be relatedby the following equations, For gas wells
P2 Ws

However, the major value of the analysis will probablybe in obtainingvalues for D and using them to determinewhether a completioncan be improved. In calculatingpotentialbenefits to be obtainedby changingcompletiongeometry, it is necessaryto define D in terms of the factorsthat can be affectedby changesin completionprocedures. The followingcases are those MfJSt likely to be of interest. Linear Flow - Gas or Liquid For linear flow, it @ that ~= 2 1+ L1 %2, 132 L2A,2 o
q q

shown in AppendixA

(3)

*o..

If, as often is the case, only the area is changed$ > = ., A22 . . . .. O.(L. .(L.

2 %2 Ir~these equations, f3= turbulencefactor,ft-l L = flow path length,ft A = crclss-scwtional area open to linear flow Subscripts1 and 2 refer to conditionsbefore and after the change ti flow geometry. Radial Flow-- Gas or Liquid

- Pwf2 =
c+D@
q

~ For oil wells


P Ws - Pwf q =

, 0

(1)

C+m

(2)

where ~s . averageformationpressure,psi

Radial flow is,the base case for determination of completioneffectiveness in ordinary wells. The followingequationis developedin

SPW Al??
e. w -,,

T.. f%..T(3NES. E. M. RTfNINT . and 0. H. GLAZE


-. -. -.. -. . . . - -.. . ---. --. <

AppendixA.

is no reason for a workoveror stimulation. However,it illustratesthe type of test results obtainedfor wells that are good candidatesfor 2r improvement. Wheneverthe slope of the plot is $ %2 W2 ~= lage in comparisonwith the intercept,the effi ciency of t e completionis suspec . 2 b !!= ?t~e ~hp12>, , ...0..(5) value of llp/Qg at npximunAp2 (AP Pws ) Is mr?e than two or three times as large as the value at the intercept,the indicationis that where h is the length of the completedinterval area open to flow near or at the well is smaller than desirable. For certain completions,such andrw?s thewell radius. Ifonlythecompleas inside casing gravel packs or fracturedwells tion length is altered, in thin producingformations,this is expected. ~2 Then it is necessaryto use @s. 3 snd4to check whetherperformanceis acceptable. InD, P2 = ~. ** , , .* (6) ordinarycased and perforatedwells, a large 2 slope is not expected,and its occurrenceindi%] cates that the number of effectiveperforations is small. Such a completionhas excess pressure EXAMPLES drop from both turbulenceend a large skin effect. Improvementwill reduce both the slope Example 1. Analvsisof Gas-Well end the intercept,and the ir~crease in producFlow Test Data tivity can be dramatic.
q q q q q

Cullender~sisochronal. flow test data for WellNo. 1 are given in Table 2? AppendixB. The results are plotted on Fig. 2. The slope, D, is approximately the same for each isochronal set, with a value of 1.70 x 10-10. The C value varies from 0.C027 for the 0.2-hourisochronal set to 0.00775 for the 24-hour set. The ~tabilizedWell Test Data increasein C is due to the chsnge in drtiage radius with time, and C will. continueto increas Data from Cullender?sWall No. 5 =e given until the well is stabilized. The value of D the expected range, and in Table 1, AppendixB. The data are plotted on for this well i.s within Fig. 1, The points all fallen the straigb.t reperforatirig probabSywill riotcause a drsmatic line except for the lowestrate, where an error chsnge in productivity. of two-tenthsof & pound in reading absolute pressurewill accountfor the deviation. The fact that D can be obtainedirom a series of very short isochronaltests is signifi From the slope of the line, D = 1.26 x cant. The time requiredfor pressurerecovery The ~ter a l~or 15-~nute flow test isor~~ily 10-10, and from the intercept, C =0.CXXM. low value of C is ~indicative of = exceptionally not long; perhaps sri hour or two. This provides high value of kh for this particularwell. The the possibilityfor rapid snd Mexpensive D value of 1.26 x 10-10 is quite high for such determination of nesr-wellflow restrictions, a low vaiue of C and probablyindicatesthat even in wells that will not stabiliserapidly. flow into the we?llbore is either ttiough a small For example,for hy&aulically fractured. wells, number of perforations or a short producing almst all pressuredrop from turbulencewil.lbe interval. in the fracturesend perforations. Flow restric tions in these locationscsnbe pinpointedby - analysisof short isochronaltests, even though With C arid D known, a true calculatedopen flow potentialcan be determinedwith Eq. 1. the formationflow characteristics are not knowr For en averagereservoirpressureof 439.0 psi, and could be obtainedonly with a long-term buildup t+st. d = 0.00028+ 1.26x 10-10Qg 9 Analysisof Back PressureTest Data Q9 where Qg is now the calculatedopen flow ~tentiaIof 3E.OMMscf/D S~nce Well No. 5 has a large calculated absoluteopen-flowpotential,there probably Cullenderalso reportedbackpressure test data for Well No. 1. It is usef%l to analyze these data SQ as to ilJ.usirate the ptential pitfallsin using backpressuretest data on unstabilize dwells.The data are given in

Either a sequenceof completelystabilized backpressuretests .m a seriesof isochronal tests provide thf? ideal data for evaluationof gas wells. Cullendertsexcellentpaperl~ protides the data used here to illustratethe analysistechniquefor ?P~~~~~~s~~e~~~~R& lized wells. Fetkovich data for oil wells.

IsochronalFlow Test Data

USE OF SHORT-TERMMULTIPLERATE FLOW TESTS TO l?UZDICT PERFORMANCEOF WELLS HAVING TURBULENCE Table 3, AppendixB. Each of the data points is for a 24-hourflow period. The first points of each backpressuretest sequencemade up the 24hour data of Fig. 2. The backpressuretest results =e plotted on Fig. 3. As expected, the slope,D, is greaterthan the slope from isochronaltests when the backpressuretest sequenceis run with the low rate first end is lower when the backpressure test sequenceis run with the highestrate first. Both of these variationsresult from the increasein drainage radius with time h an unstabilizedwell. Examle 2. Analysisof High Rate Oil-WellFlow Tests A new oil discoverywas tested at high productionrates through a limited completion interval. Prwsuxe &op from turbulencewas evidentbecause the productivityindex (PI) declinedasthe rate increased. The problem is to calculatepressuredrawdown at higher rates and to estimatePI if the perforatedintervalis increased. The well was tested. at three differentflow rates. Test data are give~l in Table 4 of AppendixB. Since the producin~zone is extremelypermeable,the well stabilizedso rapidly that no transientcould be observed. All test data we.)e taken above the bubble-point pressure,~lhich is less than 4,000 psi. Initial shut-inprcssu??e was 5$94$ PSiC

SPE 6133

At a rate of 20,~0 B/D, PI would equal 51 and drawdownwould be 392 psi.


Since C Z O, the adjustmentin PI caused by extendingthe length of the perforatedinterval can be estimatedas follows:

~ 2.
P12 ~ z% 2 ~* *(7 D, P2

Therefore,PI in this well cen possiblybe quadrupled by doublingthe length of the perforated interval. The large D value in this case probably indicatesinadequateperforationsin the sense that most of the effect of turbulenceoccurs at the perforations. Therefore,reperforattig the came intervalor doubtig the perforation density to 8 shots/f t may also increasethe PI . by a fector of 4. DISCUSSION

The analysisprocedurepresentedhere allows predictionof well performanceand diagnosis of meny well completionproblemswith kexpensive short-termtests. The procedureis particularlyapplicableto gas wells in good formationswhere completionproblemsare apt to occur. The results allow separationof mechanical problems,such as too few open perforations or too narrow fractures,from formationdamage The results are plotted on Fig. 4. The or low permeability. This permits the engineer infinitePI a+.a rate of 1,446 B/D merely to choose the type of stimulation or workover indicatesthat the preseuredrop was too small most likely to solve the existingproblem. The to read on a 10,000-psigauge. From Fig. 4t abilityto pinpointreasons for low productivity the lami.nar. flow coefficient, C, in Eq. 2 is shouldmake the test procedurea valuablenew nesrly zero? end the turbulencecoefficient D? engineeringtool. is 9.8 X 10-7. The laminarflow coefficienthas to be greater than zero, but in this case, kh is ACKNOWLEDGMENT apparentlyso great that pressure drop from laminar flow is negligible. The authorswish to expressappreciation to Mobil Research and DevelopmentCorp. and to Pressuredrawdownat higher rates can be Mohll Oil Corp. for permissionto publish this calculatedwith Eq. 2 or by extrapolating the paper. strtightline of Fi . 4 to the desiredrate to obtain a value for ?p REFERENCES alsO equals I/PI. At a ra~~ of~$~w~~ch .
P Ws Pwf F

~q= 9.7 x
= 0.0294

10-7 x 30,000

and
PI = 1/0.0294 =

34.

Drawdownrequiredto attain 30,000 B/D, therefore, is


AP c 30)000/34 = 882 psi
.

the Coefficient 1. Geertsma,J.: ~tEstimating of InertialResistancein Fluid Flow ThroughPorous Media, rSoc. Pet. EnR. J. (Oct. 1974)No. 5, 445. 2. Phipps,S. C. end Khalil,J. N.: rA Method For Determir@g the lhqwnentValue in a Forchheimer-Type Flow E@ation, f J. Pet. Tech. (July 1975) 883. 3. ~h, R; V.: Unsteady-State Gas Flow Into Gas Wells, Trans.,AIME (1961) 1151. 4. Ramey, H. J.~ Jr.: ~Won-DsrcyFlow and WellboreStorage Effects in PressureBuildup and Drawdown of Gas Wellst J. pets

SPE 6133

L. G. JONES, E. M. BLOUWT, and O. H. GLAZE

Tech. (Feb. 1965) 223. D. L., et al.: Handbook of Natural 5* ~, %%%%%%%:-Hi 00 O 6, Forckeimer, Ph.: Zeits V. deutschIng.
(1901) ~, 7* 17$2. -

7%

gas gravity (air = 1.0)

This csn be written

$
C

-C+

R9,

. . . . . . . . .(A-2

Cooke. Claude: Conductivity of Fracture Proppbts in MultipleLayers, J. Pet. Tech. (Sept. 1973) 1106c L. G. and Watts, J. W.: Esti~ 8. ~st mattig Skin EffeJt in a l%xt?ially Completed DamagedWell, J. pets Te~h* (Feb~
1971) 249* 9.

9 where the dsrcy or laminarflow term is given by . . . . . . . . . . (A-3 . . and the turbulenceflow term is
1.247 Ix$ A2 X 10 -16 EIY ZTL

. _

Jones, L. G. and Slusser, M. L.: The Estimationof Productivity Loss Caused by Perforations - ln~ludingPartialCompletion and FormationDmage, paper SPE 4798 presentedat SPE-Jit~A9thAnnu~l Fall Meeting,Houston,Tex., Oct. 6-9, 1974. IsochronalPer10. Cullender,M. H.: ~lThe, formanceMethod of Determiningthe Flow Characteristics of Gas Wells, Trans.! AIm (1955)*, 137. Testing 11. FetkovichtM. J.: tThe.Isochronal of Oil Wells, paper SPF 4529 presented at SPIMIME 48 th Annual Fall Meeting,Las Vegas, Nev., Sept. 3Wct. 3! 1973* APPENDIXA I)evelopment and Discussionof Equations

QgO.

. (A-4

A plot of 42/g vs Qghas a slope of D, and llm ~ . Thus, if most of the presc Q~O Q 9 sure drop OCC-WS in a linear geometry,the dsrcy end turbulenceflow terms can be determinedso long as L is known. , The above analysis shouldbe particularly useful for completionswith tiside casaing gravel packs. Since L, p, andk are fairly well lmown for that case, the area, A, canbe calculated. From A, the number of effectiveperforationscar oedetermined. This will allow evaluationof variousperforattigprocedures.

In calculatingpotentialbenefitsto be obtainedby changingthe completiongeometry?a new turbulencecoefficientcan be estimatedfrom Linesr Flow - Gas the previouslymeasuredvalue bytaking ~ ratio. Thus, Linear gas flow is a major contributorto pressuredrop for wells prcducingthroughinside 2 castig gravel packs or fractures. FcIreither 1 %LlA2 . (A-$ case, pressuredrop from turbulenceis fre~ 62~2A12. ..c quentlythe major factor in limitingproduction rate. If, as often ia the case, only the area is For linear gas flow through a porous media chsngeri~ of length L,5 2 2 @l= 2 8.93zTPQL , (A-6 ,* . . *
P Ws Pwf2 =
q q q q q q q q q

kA

2
-16
OZTy

A,2

This anslysis, also csn be appliedto frs.cLQg2 tured wells in relativelytight formations. In + 9 that case, L and A2 are more or less tiseparable A2 Probablythe best tests to performwould be (A-1) short-termisochronalflow tests at substan*..*. .***** ****** tially differentrates. If the twbtience flow gas compressibility where Z . dimensionless term is large, the obviouscorrectiveprocedure coefficient to use for obtaininghighe ~ flow rates is tO *R T= reservoirtemperature, widen the fracturesinceA appearsin the denominatorof-the turbulenceflGw term. We M = viscosity,Cp L = flow path length,ft suspechthab pressuredrop from turbulenceis a k = permeability, md major factor in limitingflow rates from f.acerea open to linear A= ;ross-sectonal tured wells when the fracturedformationis flow, ft2 relativelythin. Based upon evaluationof a number of frac~uredwells, the fracturedesigns @ . turbulencefactor, ft-l
1.247

X 10

USE OF SH~T-TIRM MULTIPLERATE l?WW TFSTS TO FREDICT PEWORMANCE OF WELIS HAVING TURBUIXNCE could be alteredto take adventageof optimum flow conditions. Also, h those cases where turbulenceis obviouslya major problem in limitingflow rate, restimulation jobs producing greaterfracturewidths might be considered.

SPE 6133
1

term i% written above as though the whole intervalwas completed. This is seldomthe case end it is useful to considerwhat happensto the turbulenceterm in a partiallycompletedwell. In that case, flow convergesfrom the entire producingintervalinto the completedinterval. In most cases the last few feet of flow paths Radial Flow -Gas are essentiallyradiel.~~9 Also, most of the The base case for determination of compl.e- turbulencepressurei~op takes place in this tion effectiveness for ordinarywells is radiel last few feet of the flow path since the area perpendicular to flow becomee very smell in that flaw. The equationdescribingradiel zone. Therefore,for practicalpuzpses, the semisteady-stat.+ gas flow is4 value of h h the turbulenceterm can be re1.424MzTQ placed by the length of the completedzone, ~. 2 q (ln 0.472> Then, P Ws - pwf2 kh
, s, ,

3*16X
q

:,f%z!(i-t),~
****9*
q

, 3*MX
. %
q 9*****

10 8 ~ 2 P

~yZT -. r

~ Q9 w

.**.*.
= = = =

***-

*.***,

**e

(A-1O)

where re rw h s

externaldrainageradiusf ft well radius, ft producingformationthickness,ft skin effect excludingturbulence effects

3.I6 X 1018 f3~ZT


snd D= .. hp2 rW
q

A-l)

Eq. A-7 can also be rewrittenas

!$ = C+DQ
9
1.424vzT c= -7T7

9**
is

(A-2)

where now the dsrcy term


(In

0.472

~ ~ w

This formulationfor D has considerable @plicat,ions concerningwell test results. The vahe of D will shsnge each time somethg is done to the weLl ~hat changes the flow patternsinto the wellbore. The effectson the turbulencecoefficientof changingcompletionlength,or well radius,w alteringthe formationcan be estimated by comparingbefore eni after completion conditions. Thus?
~

S)

(A-8) ~=

and, since I/re normally is very smsll, the turbulenceterm is


3.I6 X 10-8
lx+
q *****S

1%22
rw2

. . (A-12)

BY ZT

~h2ro 2 P] WI

1 TQ9
d*

= h2
q S*****

If only the completionlength is altered,as often will be the case,


(A-9)
h 5=P2 z P1

A plot ofAp2/Qg vsQg again has a slope of D,andC=


lim Q~O A& Q 9

(A-13)
q

Linear Flow - Oil

Linear oil flow also has a darcy pressure drop and a turbulencepressuredrop. In many cases the turbulenceterm is quite low end can be neglected. However,for high rate wells in Correlationsare availablefor expected naturallyfractkmedzones,or for inside casing values of 19for given values of k,5f7 Occasion- gravel-packed completions, the linear flow ally tha IQg term will be orders of magnitude turbulenc$componentof pressuredrop can become higher then expectedfor the known formationkh. the major factor M I.imiiing flow capacity. In that case, it is obvious that the completion is inefficientend that more end deeper perfora- ++Note here that o~ def~tiOn of D is related tions are needed or that the completioninterval to that of Rsmey4 by the equation shculdbe lengthened. D =-r~, where ~is Rameyrs value. PartialCompletionEifects- The turbulence

TheAp2/Qg vs Q plot allows differentiation between smple Samage end problems invo~vingthe turbulenceterm for radial flow.

sPl? -. - 61~~ . -==

L. M. BWUNT. and 0. H, GLAZE .-.G. J-. . . ...E. ---appreciable, the well probably shouldbe stimulated or reperforatedor the completedinterval shouldbe extended. For radSsl flow,
pq2B2p *
q

cm

For liquid flow, Forchheimerts equation6 be written


dJ?=
qMB

1.127 x 10-3kA 9.o8 X 10-3


+ A2 A-u)

~= 1.127X

qvB
10-3kA Bq2B2p
1

Therefore,for linear
P2 -PI= 1.127

flow,
qpBL X 10-3kA

9.08 X 10-3
+

A2
q PB

. (A-19)

where A = 2firh. This is easyto integratesnd P2

where p2 = PI = L . A = :. = p .

13 fjq2~2~ L 9.08 X 10 + ,(A-15) A2 pressure at the entranceof the linear flow path, psi pressure at the efit of the flow pab, psi len@h of the flow path, ft cr -sectional area perpendicular t I1OW, fta flowrate, STB/D formationvolume factor,RB/STB fluid density,lb/cu ft

-p,=
1.127

in~ X 10-3( 2mkh)

2
I

9.08 X 10-13

@q2B2p

**.*.*

4 n2 h2 .*****
q

()
I -
q

I 1 (A-20)

2 ***

This can be written as

where p2 is the pressure at some outer ratiust r2, and pl is the pressure at the inner radius$ rl. Ifwe letp2=pws, r2=re~ P1 =Pw~ ~d rl = rw; if we account for skin effect and pseudosteady-state flow in the laminsrflow term; and if we assume re is fairly large, then qVB
P Ws Pwf =

&
o
1

= C+Dq,
=
9.08x
p BL

. . . . . . . . . . (A-16
q

where c

***

(A-17
.

1.127x
( in \

10-3(2~kh)

1.127+

lO-3kA
10-3 6B2PL

0.472 ~~s
r ./

)
I

and

D =
Iim & ~

,*,

* (A-la
9.08x
+ iO

13 9q2B2~,

A plot ofAp/q vs q fives the laminarflow term


C, as C s q-o ~

4 TI 2h2 and the turbulenceflow once again we obtti the form

()

. .

(A-21:

r w

term, D, as the slope. Except for oil wells with exceptionally high flow rates, the turbulenceflow term shovldbe negligiblewhen comparedwith the laminar flow term. If this is not the case, t~ completiondesign shouldbe reviewedcarefully because the flow srea is probablytoo small. Even though the turbulenceterm is not a large percentageof the total pressuredrop for sn oil well, its noticeablepresenceindicates that the laminsrflow term can-alsobe dropped considerably by providingmore srea for flow into the wellbore. Radial Flow - Oil

= c+.Dq , . . . . . . . . . (A-16: q where C, the laminarflow coefficient, is given , by


PB c=

1.127 X 10-3( 2nkh)

inO.472

~+s rw )

**

(A-22

and the turbulencecoefficientis


9.08x on 13 fjB2p 10 St c(A-23;

The turbulenceterm in radial oil flow shouldbe negligiblefnr most oil wells. If the term is measurableand if pressuredrop is

b m2h2rw

USE OF SHORT-TEH$l MULTIPLERATE FK)W TESTS TO PREDICT. PERFORMANCEOF WELIS HAVING TUR13ULRiCE Once again, a plot of ~p/q vs q has a slope of D,snd C=
Iim &

SPE 6133

q-o

partialCom@etion Effects-It is worth noting here that D sgsin is very much a fun,ction of the length of the completedinterval.. If oriy psrt of the interval. is comp~eted,h is replacedby ~ and
9.o8x DE 1013 BB2P * (A-24)

Again, a comparisonof completionsskmild tielude an adjustmentfor the lengthof the completion. Thus, assumingeverythingelse is constsnt,for two differentcompletion lengths ~2 >=~* . 2 h
P]

(A-25)

~n2h2r ~ .b*

APPENDIX WELL TEST

DATA FOR EXAMPLES

Table 1 Performance Data of Gas Well

No.

Shutin Pressure psia

Flew Rate MSCF/D

2 P Ws -

$2! Pw; q

439.6 439.8 439.0 439.0 439.0

2,231 4,841 8,373 12,484 16,817

1,010 4,320 11,270 23,250 40,600

.000453 .000892 .00134 .00186 .00241

Table

2 No. I

Isochronal Oata of Gas Well

Isochronal

0.2 X2 x

hr.

data

Isochronal Q X 10_3 ~~x

1 hr. @

data &2/~ .

QO X 10=3 2009 299~ 4130 7327

lo=3

AP2/Qa .00283 .00307 .00308 .00356

5.68 9.21 12.72 26.09

1994 2937 2941 4052 46s6 7092

8.26 12.87 13.28 18.40 22.52 35.52

.00414 .00438 .ool+52 .00454 .00484 .00500

3 hr. 1980 2905 4587 6887

data

24

hr.

data

10.38 15.72 27.26 42.99

.00524 .00541 .00594 .00624

1447 4440 9900

14.56 38.67 97.70

.00748 .00871 .00987

Back

Pressure Q
X

Table Test

3 Data

on Gas 2x

Well 10-3

No.

I AP2/Q

Date

10-3

10-3-44

9900 7091 4360

97.70 70;73 46.16

.00987
q

00997

.01059

10-24-k4

4440 6$)82

38.67 79.17 $2.35

.0087

.0108 .0112

8212

12-11745

1947 2841 3941 5165

14.56 25.07 38.82 50.53

.00748 .00882 .00985 .00978

Performance

Data

Table for

4 High

Rate

Oil

Well

Rate

Bottom Hole Pressure

& @@ PI

1446 6199 8115 20,(100* 30,000*

5448 5410 5383 5056* 4566*

0 38 65 392* 882k

0 .00613 .008 .0196* .029*

0 163 125 5 I* 34.0*

Calculated

Values

.0026 n .0024 / .0022 1 .0020

.0010

.0016

.0014 da m

Go

2 .W1 A

/~

.0010

.000 $

.0(0

.00 06

.009
.0004

.0 c412-

.008
2 4 6 Slo Qg Z 10-* 12 14 16 [e

FIG,

1 - ANALYSIS OF STABILIZED WELL TEST DATA, WELL NO. s,

0 #.004

/
/

/
.003 .002

024 HI A 3HI o I HI u .2HI

.001

6678

FIG,

2 - ANALYSIS OF ISOCHRONAL FLOW TEST DATA, WELL Nfi, 1,

. . .

. .o~o

... .

.018

.016

.014

.012

. -- = .O1o 40

.Oos

0 .00 6 4

.00 4

.00 2-

00
QQ E 10-8

FIG,

3 - ANALYSIS OF BACK ?RESSURE TEST, WELL NO, 1,

.020 .016 .016

.012

.00B

.00 2 0 2000 4000 6000 8000 10,000 q(bbl/doy) 12,000


I 4,000 16,000 18,000 20,000

FIG,

4-

ANALYSIS OF HIGH RATE OIL WELL FLOW TESTS,

You might also like