Spe 28306 Ms Hole Clean
Spe 28306 Ms Hole Clean
Spe 28306 Ms Hole Clean
Society of PetroleumEngineers
SPE 28306
A
Copyright 1994, Society of Petroleum Engineers, Inc.
Tfds paper was prepared for presenta$on at tha SPE 69th Annual Tgchniml Conference and Exhlbltion hdd in Now ohms, LA, U. S.A., 25+8 .Septemb.ar 1994,
This papw was SaIected for presentation by an ePE Program Curnmittee following review of lnfomliai cantdnad In an aksfracf silbmitted by the authm($~ contents of the paper,
as presents-d, have nof bee reviewed by the !Mety of Petroleum Engineers and am subjwt to correction by the author(s). The material, as presented, does not necessarily reflect
any position of the Sodety of Petroleum Engineers, its officers, or members. Papers presented af SPE meetings are subject 10 publication re.dew by Ed[torial Committees of the Sodety
of Petmle.m Engleers, permission to copy Is restricted to an abstract of net ore than SCU words. IIlusttafions may not be copied. The abstract sfm.ld wn!ah conspicuous .Mmwfedgmem
of wher8 and by whom the paper Is preswtod, write Llbrarlan, SPE, P,O. Box 632s!36, Richardson, TX 7503S-3336, U .e,A, Telex, 183245 SPEUT,
139
,
2 A MECHANISTIC MODEL FOR CUTTINGS TRANSPORT SPE 28306
wellbore angle.la The models are accessible via main frame or builds, the mud velocity over the bed increases. The cuttings
personal computers end uee readily available data as input buildup process continues until the mud veloeity over the beds
parameters. surface eventually reaches the critical value. At that condition,
Laboratory experience indicates that the flow rate, if high the bed height remains unchanged. If additional cuttings axe
enough, will always remove the cuttings for any fluid, hole size, deposited onthebed, themudvelocityin theneighborhoodofthat
and hole angle. Unfortunately, flow rates high enough to region exceeds the critical velocity As aresr.dt,the stronger fluid
transport cuttings up and out oftheennuluseffectively cannotbe forces will dislodge the protruding outtinge. After these extra
used in my welis due to limited pump capacity and/or high cuttings are moved downstream, the local equilibrium bed height
eurfeee ordownhole dynamic pressures. Thw ie particularly true is then re-established. Thus, the equilibrium bed height is
for high angles with hole @eZ@ger then 12?4 in. High rotary formulated as a function of the critical mud veloci~.
speeds and backreamin g are often used when flow rate does not Furthermore, the critical transport velocity is the critical
suffice. velocity that gives a zero cuttings bed height.
Drilling fluid rheology plays en important role, _although
often there are exmflicting statements as to whether the mud
should be thick or thin for effective transport. It is common when
drilling high-angle wells for elevated low shem-rate theologies to
be spec~led. Ty@callx the Fmm 6-rpm and 3-rpm dial readings
are set at some level thought to aid in hole cleaning at high angles. 7<
1
Wellbore
Many settheselowshearreadingz (inlbf/100ft2)equivalenttothe
hole sizeininches. Thisrecommendationas wellazotherrulesof U(
thumb have been presented by Zamora and Henson.18~17 -.
Cuttings J
The model described below was developed to allow a
complete cuttings transport analysis for the entire well, from n
surface to the bit. The mechanisms which dominate within Mud velocity profile
(x and z components) s
dflerent ranges of weilboreengle areuzed to predict cuttings bed
heights and ,=ukir cuttimgs concentrations as functions of
operating parameters (flow rate, penetration rate), wellbore
conf@ration (depth, hole angle, hole sizeorcasinglD, pipe size),
fluid properties (density, rheology), and cuttings characteristics
(density size, bed porosity angle of repose), Parameters that are
not currently taken iriti account include pipe eccentricity and
rotary speed.
This paper has three major section% (1) the fwst identifies
the modes of trmmport and outlines the mathematical
development of the model, some of which is given intheappendix; Formation -%
140
SPE 28306 R, H. CLARK AND K L. BICRHAM 3
derived for the three patterns. However, the governing end the muds rheology is assumed to be governed by the
mechanism is the one which dominates the flow at a particular Herschel-BuIkley viscosity law. The static forces are the
wellbore angle. Two mechanisms are based on the forces required buoyancy force, F~ gravity force, Fg, and the plastic force, FP,
to displace asingleprotrudingcuttingfrombeds surf-, namely which is due to the yield stress of the mud. The dynamic forces are
these equations calculate the velocities that causes acuttingto be the dragforce, FD,Iiftforce, FL, and pressure gradient force, FAP
either rolled or lifted from its resting place. The third equation is Theysxeallzssumedtoaot throughthecenterofgravity. The mud
basedonthe Kelvin-Hehnholtz stabiMyofthemudlay erflowing circulation rata is held constant.
over the fluidizedbed. Finally the fourth equation is baeedonthe
settling velocity of the cuttings, that is, the annular velocity RollingMechaniem. Forthecsaeofrolling, themomentsdueto
reqtied to limit the suspended cuttings concentration to five forces are summed around the support point, a(x,z); nmnel~
percent by volume in the flowing mud stream. IxI(F. +FJ+ lz\@. -Fp)+g(F, -F,)=O (1)
Equilibrium Cuttinge Bed Height Models. Figure 2 shows a
where the length of the moment arm for the buoyancy and gravity
stationary cuttingebedthathes formedonthelower wellbore wall forces is
in an inclined well with a wellbore angle, a. At high wellbore
anglee where the wellbores complementary engleis less than the t = Izl(sina + cosa/ten@) (2)
cuttingsengle ofrepose,~, astationarycuttings bed accumulates
Moreoveq O <as 90; likewise, O s @ s 90. Ftily the figure
in the lower part of the wellbore cross section. When the wellbore
shows
complemental angle, 90- a, is Iesethsm$, the outtinghae to be
either rolled or lifted from the bed surface in order to move. @ = arctan(z/x). (3)
Suppose that the cuttings bed height is in equilibrium with the
When the dynamic forces exceed the static forcee, the cuttings
prevailing conditions. If the dynamic forces acting on the
tend to roll along the bed in a moving cutting zone. The dynamic
stationmy cutting can be calculated as a fmction of local mud
forces generally increase with mud velocity. Exceptions may be
velocity, U, then the mud circulation rate needed to dislodge the
possible; for example, Coleman20 and Davies and Szmad2s
cutting can be determined. This notion is en exteneion of work in
experimentally observed that the lift force is negative in a smell
other areas, such es sedimentation,18~Ig,zo soil erosion,zl and
range near a particle Reynolds number of 100 and is positive
slurry transport.zz
elsewhere.
F~ FP+(Fb-F.Jeina = O. (4)
141
-. . ...
,
4 A MECHANISTIC MODEL FOR CUTTINGS TRANSPORT SPE 28306
F~ = C~~@J2, nda
(5) F& = r~, (lo)
the lift force, where
FL = C@ QU9, (6)
(11)
the buoyancy force,
where Dhyd is the hydraulic diameter of the flow area (ses
Fb = gQ~, (7) Equation (16) and Equation (17)), P is the pressure, z~is the wd
and the gravity force, shear stress, end ~ is the mud yield stress.
F. = gQ=~ (8)
Rolling and Lifting Bed Height Equations. Two equations
where CDis the drag coefficient, CLis the lift coefficient, and g is forcriticrdvelocity may beobtainedbysubstitutingEquations (2)
the gravitational constant. end (3) end the ancillary equations (Equations (5)-(11)) into
The following two equations are derived in the appendix Equations (1) and (4). At high wellbore angles, one of these
(Equation (A-1) end Equation (A-4)). The plastic force, resulting equations for critical velocity may govern the flow. For
Fp = =[$ + (Jx/2 - @) Sill@ - cos@sin@], the case of rolling, the geverningequation for the critical velocity
(9)
is
1/2
1
4[3~@ + (rc/2 @) SinzI$ - COS!$Sill@)&n@ + dg(Qc Q)(cosa + 5~ati@) drl
u= 3Q(C~ + CL tan@)
(12)
- [.,
For the csae of lifting, the governing equation for the critical veloci~ is
.. . .
1/2
4[3~@ + (x/2 - @) Sinz@ cos @SkI@) + dg(Qc - Q)Sins]
-u =
..
[
,,-
3QCL
-.. .
1 (13)
Both Equation (12) and Equation (13) give a value for the Coneiderthestratifiedflow arrangement ehowninFigure 3.
critical velocity of a cutting. The velocities calculated by these There is a nearly cuttings-flee mud layer flowing over a mostly
equations are the undisturbed velocities, that is, the axial velocity cuttinge-fiiled fluid layer. A smell-amplitude wave propagates at
acting above the cuttings bed at a point that would be occupied by the inter&e as long as the flowing conditions are stable. The
the cuttings center ifit were in place. These equations calculate inviscidKeIvin Helmholtz stability theory provides amethodfor
the velocities that would either roll or lift the cutting from its predicting the onset of unstable conditions between inertial and
resting place. In general, these two mkulated vekes will be gravitational forces acting on the interface,z5 that is, the value of
different. In such cases, the lower value will be the dominent one mud velocity that causes the lower layer to disperse cuttings
providing that other conditions cmemet. throughout the entire cross section. Wallis end Dobson20 give a
clear description of the instabili@ condition for stratified
Kelvin-Helmholtz Stability Model. When the behavior of gas-liquid flow. Their result is adopted here as follows:
the cuttings is observedat low wellbore angles intheflowloop, the 1/2
nature of the mud and cuttings slurry is a churning motion. The Dqg(Qb Q) sina d e
TJmk > (14)
process is reminiscent of the behavior of a gas-liquid flow when Q l~1w ( )]]
[
its flow pattern is changing from stratfled to slug flow. The
app~ce Oftie fluidized bed is similar to the liquid layer, and
the mud layer flowingover the bed behaves like the gas layer. The @b = Q.(1 (15)
interface between these layers has a wavy churning nature. Dq is the equivalent diameter of the area open to flow, end ~b is
Occasionally, wisps of cuttings are swept up into the mud layer. the bed porosity. Sometimes, Qbis called the submerged bulk
There, they are carried downstiefi and settle back into the density Whentheaverage mixturevelocity, Um~, intheopenarea
fluidized bed. The process is persistmt, and it appears to be above the bed exceeds the RHS of Equation (14), the interface
random. betweenthelayers is unstable. The minimum transition velocity
is when U~ equals the RHS of Equation (14).
142
w SPE 28306 R. K. CLARK ANf) K L. BICl@A.M 5
% = m ()
The plug diemeter ratio is
143
.
6 A MECHANISTIC MODEL FOR CUTTINGS TRANSPORT SPE 28306
Cuttings Concentration. The feed concentration isdefmedas relationship, are all used to determine the critical velocity.
follows: However, the predicted velocities from both Equations (12) end
(13) muetbe put onanequivalentbmiswiththosepredictidfiom
(23)
0 & Equations (14)and(33),nsmely theaveregemixturevelocity. The
The average concentration, c, of cuttings in a short segment with critical velocity is determined according to the following
length, Az, and cross-sectional are+ ~ can ha calculated as 1. For near-vertical cases, when the values calculated by
follows Equations (13) and (14) are less than the one fkom
c =c.(1~)+cJ:. w) Equation (33), the critical velocity equals Equation (33)s
value. Ifthecirculation rateexccedsthisvalue, the suspended
The cuttings concentrations in the plug and annular regions are
cuttings concentration will remain less than five percent.
assumed equal. This means that the suspended cuttings are
However, if the mud circulation velocity is less than the
uniformly distributedacroes the ereaopentoflow. Obviously, this
cuttings settlingvelocity, the cuttings willeventuallybuildup
has a major impact, and it probably is a function of wellbore
in the wellbore and plug it
geometry, mud properties, cuttings properties, and operating
2. Forlow-angle cases, where the wellbore complementaxyangle
conditions. It could stand alone as a research topic. Thus, we
is greater then the cuttings angle of repose, the remdts of
obtain
Equations (12), (13), (14), and (33) are ranked emsllest to
u =_.cu.(l c)
.. (25) largest.
c-c.
3. For high-angle cases, where the cuttings angle of repose is
where
greater than the wellbore complementary angle, the restdtsof
u. = U=(I ~) uq~ (26) Equations (12), (13), and (33) are ranked smallest to largest.
is the average settling velocity in the axial direction. The The critical velocity equals the fwst value that exceeds
components of the settling velocities (see appendiz) in the axial Equation (33)svalue. If~here=e none, thenthecriticslvelue
direction are equ-alsthe one calculated from Equation (33).
_ 1/2 compared with date from flow loop experiments conducted over
three separate two-week intervals. The first two sets of tests were
1
4dg(QC @
(29) conducted on the 5-in. flow loop at the University of Tulsa. This
u: [
3QCD
equipment has been described extensively by others.3-6 The
1[2
third set oftests was conductedontheUniversity ofTtiasnewer
4 dg(e. - Q)
u~ = iccy cosa, (30) 8-in. flow 100P.7ZS
~3
[{ 1]
Fluids used in the experiments coneistedof water, solutions
dUQ of HEC, xanthan gum and PHPA in fresh water, end bentonite
ReP = (31)
w. slurries before endefteraddition of en exteudmgpolymer. In all,
3ty 158 tests were run on the 5-in. loop end 60 on the 8-in. loop. The
(32)
s = dg(@c Q) tests were run at angles ranging from neer-verticel
CD is the drag coefficient of a sphere, ~ is the yield stress of the (200 minimum) to horizontal (900). The inner pipe in the 5-in.
mud, end Vais the apparent visc6ii@ of the mud at a sheer rate loop was set both concentric end eccentric with the pipe %-in.
resulting from the settling cutting. above the low side of the annulus. The pipe wee concentric in the
The value calculated using Equation (25) is the minimum 8-in. loop tests. Pipe was only rotated in the 5-in. loop tests.
acceptable mixture velocity requiredforacuttingsconcentration, Two types of test results were obtained (1) a visually
c. Plgott recommended that the concentration of suspended determined critical flow rate and (2) the equilibrium annular
cuttings be a value less than five percent.l With this limit cuttings concentration as a function offlow rate. The criticelflow
(C = 0.05),
Equation (25) becomes rate wee taken to be that at wbichnocuttingsbed was formed; i.e.,
all cuttimgs were observed to be moving upward cind no
u mix ==- (33)
0.05- 0 accumulation of cuttings was occurring. The cuttings
where ~ <0.05. This implies that the penetration rate must be concentrationiiithe ennulusisessentially equal totht oftbcfeed
limited h a rate that satisfies this equality. under these conditions. As the flowrate is lowered below critical,
Equations (12) fid(13),thecriticrdvelocityrelationshipsfor cuttings begin to accumulate and form either a moving, churning
rollingendlifting, Equation (14), theKelvin-Helmboltz stabiii~ bed at low angles or a stationary bed at high angles. The dividing
relationship, and Equation (33), the critics3 mixture velocity singleis taken to be the complement of the angle of repose.
144
.
Foralltesta,theannular cuttingeconcentrationwasallowed measured data for the large cuttings (0.43-in.) and the model
to reach a steady state, cuttings injection wqs stopped, and the predictions. The quantitative egreement is not so good for the
cuttinge were flushed out of the anmdue end weighed. From the small cuttings, eithough qusditatively the chsnge in cuttings
cuttinge weight and density a volume percent concentration in
the annulus was calculated. This concentration could be
24
converted to a cuttings bed height knowing the cuttings bed
z-
porosity end the po~tion of the inner pipe.
23-
a x Measur6d Concentric 0.42
~ ,* -
Critical Transport Prediction. Figure 5 shows the vieually v Measured Eccentr?c 0.4S
S 16 - Predkted 0.18
determined critical flow rate, described above, as a function of PFa&h3d 0.4s
hole angle foraxanthrmgumfluid with the properties listed with
the figure. The hole and pipe size, penetration rate, and cuttings ~ 10-
sise ere also listed. The critical flow rates determined with the g -
pipe concentric and with the pipe eccentric are both indicated. 8-
v
4-
2-
Flow Rste (gpm) Mud Velocily((pm)
0
6OW1M13374O1O31SO 2W2222402E02W
0
j -
o~
# , I +
0 108?2040= ev702090
WelboreAngle, deg P 5
E
~ !0-
Mud Xanthan Gum Dsnstiy: 8.3,ppg
Pv: 3.5 Cp PiPeDiet 2.2 in. 5-
YP: 8.0 IM1OOR2 Hole Dla .5.0 In.
YZ 2.5 b/100it2 Cuffirsw
,- 0.18 Io_
ROP 50.0 fph w IDI 120 140 160 180 m 2ZU 240 =0 230
FlowRate (gpm)
I
Fig. 5 Critical transport comparison.
Fig. 7 Cuttings transport in a 5-in. flow loop at 50
145
.
40 Cultrll so
* PipePositbn Size (In %-
+ MeasuredConcentrk0.18
35- 0 Measured Eccentk 0.i8
m-
x Measured Concentrb 0.43
2A-
g ~. -
v Measured E.xan!rb 0.43 gr2 -
Predbtd 0.18
Pred!!t& 0.43
[ 25 - B
~ 18-
g 20 -
0 g; :
~ 15 -
~ 10-
g ,0 .s8-
E
0
5-
2-
0
80 IW 120 140 180 130 240 220 240 =0 280 1msN3m4cOs10 sm7c08co
Flow Rate (gpm) FlowRate (~m)
Fig. 9- Cuttings transport in a 5-in. flow loop at 90. Fig. 11- Cuttings transport at high englee in en 8-in. flow loop.
Figures 10 end 11 compexe model predictions with
FIELD APPLICATION
experirnenti data from tests on the 8-in. flow loop With an
extended bentonite mud. Agein, the highest flowrate data were The Wttings transport model, in its easy-to-use personal
taken to represent ~itical conditions. These are again computer format, has been applied to many different drilling
considerably above the model prediction. The critical flow rate situations. A number of these are discussed below.
prediction in the 8-in. loop is certainly more in line with field
DrillingLarge-Dweter Holes in Deepwater Operations.
experience then that based on visual observation. Agreement
Thefwst stringof pipe set duringdeepwater drillingoperetions is
between experimentendmodelpredictionis quitegoodforeach of
the sixhole angles. Unfortunately insuflicientdatawere takenat a.SO-in.or 36-in. structural pipejetted several hundred feet below
the lower hole anglee to eesese the model predictions fully. the mud line. The first interval drilled end cased is for either
20-in. or 26-in. caaing. This interval is usually drilled with
Inflow loop tests with water and other low-viscosity fluids,
the model consistently underpredicted the annular cuttings eeawater end viscous sweeps with mud returns to the seafloor.
concentrations at angles above 50. It appears that the fluid The large hoIe size smdlow-viscosity drilling fluid (eeawater) will
rheolo~ is given more importance in the transport model than ia result in abuildup ofcuttings in the structuraIpipe srmuluswhich
can, if the fracture gradient at the shoe of the structorcd pipe is low
usually seen in high-angle flow loop experiments.6)8
enough, result in loss offluid. This loss is one of several causes for
what ie called shallow water flow, i.e., abreekthrough offluid to
28- \, + 20. Mea%. the sea floor mound or away horn the structural pipe.
s -
\ 20 Pred.
The cuttings transport model was used to examine this
A 35.Mea,.
30. Pred. problem end to identify corrective action. Table 1 lists the
v 4W Mess.
- - . 40- Pred. predicted steady state cuttings concentration in the enmdus of a
36-in. structural pipe (34.75-in. ID) es afimction of flow rate. The
pressureat thebaseofthe 200 ftlong, 36-in. pipegeneratedby this
~ to - cuttings-laden fluid is also given. If the pressure imposed by the
30 - cuttings-laden fluid exceeds the fracture pressure at the base of
oil -
4- the 36-in pipe, fluid flow to the mud line may occur. For weak,
2- -~ shallow sediments in the deepwater Gulf of Mexico, the fracture
o
1WSM3W4WX.3 w07m8m gradient may beequivelent to only 30 or 40 psioverhydrostaticor
Flow Rate (gpm)
118 to 128 psi total.
ig. 10- Cuttings transport at low angles in en 8-in. flow loop. The two portions of thetable correepondta drillinga31%-in.
hole at 50 Whr with seawater end the use of a viscous sweep
(density = 8.9 lbm/gaI, plastic viscosity = 9 CP,yield point =
40 lbf/100 ft2, yield stress = 15 lbf/100 ft2). The cuttings bulk
density is 2.05 g/cm8 and the she ie 0.25-in. The drillpipe size is
5-in. in thie example.
146
.
Table 2
CUTTINGS LOADING IN 36-in. STRUCTURAL PIPE CUTTINGS CONCENTRATION IN A WASHOUT
F1OW DrillwithSeawater DrillwithSweep FlowRate Annular EquilibriumCuttings
Rata (gpm) Velocity Concentration(%)
(gPm) Cuttings Pres9ure cuttings Pressure (ft/min)
Concentration at Shoe Concentration at Shoe Experimental Predicted
(%) (psi) (%) (psi)
100 25.8 33.0 26.8
750 51.0 134 21.8 llz
125 32.3 24.9 21.5
1000 45.0 129 15.4 107
150 38.7 19.5 16.7
1250 39.8 124 12.3 104
1500 35.3 120 8.6 101
Experimental data from M (Reference No.-27).
1750 31.1 116 7.0 99
2000 27.2 113 4.5 97 Redevelopment Drilling. Redevelopment of axistiig fields
Pipe jetted to 200 ft balow the mudline, drilling 31%-in. hole. often involvae reentering an old well, cutting a window, and
rhilling out to a newbottomhole location. Such wells czn have
The model provides guidance on drilling the 26-in. casing compkx directional progrmns. This was the rase in a recent
interval such that SW1OWwater flow can be minimized. It is offshore well in which awindow wascut in a curved conductor, the
obvious from Table 1 that a high flow rata is essential, as are well kicked to an angle of ovar 40, droppad to near-vertical, and
periodic viscous sweeps, to keep the pressure at the base of the then turned sharply and eventually completed as a horizontal
structural pipe at a tolerable level. Drilling continuously with a wefl. During drilling of the 12]/!-in. hole at an angle near 85,
sweep would be succesefid, although the total volume of sweep problems were axperiencad on strip out of the hole at ameseured
required for drilling the 31%-in. interval may exceed the rig depthof 6710 ft (5700ft TVD). It tookexteneivebackresmingand
mixing capability. circulation to compIete the trip out of the hole successftily
The cuttings concentration levels shown in Table 1 are The output for en analysis of this situation by the cuttings
essentially unch~ged for each of the two d@rent operational trsmsport model is shown in Table 3. The input parameters
procedures in common practice in deep wate~ (1) drilling a pilot include the mud type, the rheology model chosen, the penetration
hole to the 26-in. casing point and then opening to 31Ys-in. or rate, the mud flow rate, the mud properties (density, plastic
(2) drilling a 31%-in. hole in one pass. The sane cuttings loading viscosity,yield point, endyield stress), end the cuttings proparties
will eventually occur in the 36-in. ennulus whether or not a pilot (density diameter, bed porosity, and angle of repose). The
hole is drilled before the final hole size is reached. If the cuttings measured depth, hole angle, hole size, and pipe size complete the
from the pilot hole arecleaned out of the 36-in. snnulus, they will input data required for conducting the analysis. These data are
build up again as the pilot hole is opened. Itis ilso interesting to included in the output es indicatad in Table 3. Note that 133/5-in.
note that the cuttings loading is virtually independent of easing (12.347-in. ID)hadbean set at 3010 ftmeamu-eddepth, and
penetration rates that a-e typical of deepwater operations. that 5-in. drill pipe end 180 ft of 8-in. drill collars were used.
If~Ything,the model may underpradict the magnitude of The results of the emdysis at each depth include the
the cuttings btildup, se sugges~d by comparison with the following: the mud velocity in the open area above the cuttings
experimental data of Ali shown in Table 2.27 Alis data were bed, the equivalent circrdating density (ECD), the mud pressure
generated by placinga 10-in. diameter washout, six feet in length (circulatingwithoutcuttings andtotaIwithcuttinge), thecuttings
in the verticrd 5-in. flow loop at the University of Tulsa A concentration (in the circulating mud end total in the anmdus),
Carbopol solution was used as the drilling fluid. the areaopen to flow, andtheheight of the cuttings bed. Figure 12
A similar amdysiscanalsobe-conducted to examine cuttings depicts much of the same information but in a format that allows
buildupinalsrge-diameterdri~ingrk~. Theneedforhighermud the location of cuttings accumulations in the wellbore to be more
viscosity, viscous sweeps, end/or additional flow rate by boosting readily identified.
the ricer can all be aasessed end operational practice set as The asterisk in the fa right-hand column of Table 3
necessary. Monitoring the pressure at the base of the riser is a way indicates that the cuttings accumulations at this location me in a
of assessing how effective such practices are at keeping the riser movingbed end will avakmche down the wellbore if the pumps are
clean. turned off without first circulating them out of the well. Where
there are no asterisks (depths from 6310 to 6525 ft), a stationsg.
bed three to four inches in height is predictad.
147
.
1
while drilling the 17%-kI. hole in each well, the extent of the
stationary bed wee far Iesa in the C2 well than in the C3. The
heights of the stationary beds are predictad to have been about
equal in both wells, five to six inches depending on the hole angle,
butthetotalvolome ofcuttingsinthestationarybed intheC3 well
was over four timee the volume in the stetionmy bed in the
C2 well. This reduced cuttings volume in the C2 well resulted in
am Iesstimeepent backreamingat ahighrotary speed, shout the only
.
Well Cia. E( Mu-#el. practical way cuttings can be removed in a large, high-angle hole.
m. PI
Mud 12.5 ~g ROP 50.0 @h
Each welliapredicted to have contained about the samevolumeof
Drillpip
Pv 40.0 Cp Cko. Rate 620.0 ~m cuttings in moving beds, outtinge which can be circulated out of
Wellbore v!? 17.0 Ib/looitz Cuttisgs 0.25 in.
VZ 6.0 lb/100f12 the well given eufticient circulation time.
The cuttings transport model predicts few hole cleaning
Fig. 12- Cuttings analysis in a redevelopment well. problems in the 12%-fi. end 8%-in. intervals in both wells, even
though these interva.lsweredrilledatangles of80ormore. While
Several pointa can be made flomthis analysis: (1) a buildup some problems were mentioned in the StatOilpapers, they were
of cuttiige is likelv in two intervak (z) where the hole angle ia50 not of the same magnitude se experienced in the 17Yz-in.interwd.
or less, theseouttingsarein amovingbedandcen becirculatedout One of the objectives of the well path used in the C2 well was to
of the well but will avalanche down the well if not circulated out reduee torque end drag. The cuttings treneport model indicates
firs~ (3) cuttings canied in a moving bed contribute to the total that the@eof path eelectadforthe C2 wall ie also bentilcial from
wellbore pressure (ECD); and (4) a stationary bed can exist at a hole cleaning standpoint. This has also been noted by Raei.15
angles above 50 and up hole fkom the drill collars. Table 3 and Thus, one of the uses of the cuttings transport model ie to design
Figure 12 show the situation as it occurred. The model input well paths that yield the fewest hole cleaning problems, assuming
perametera can be varied to see what action is most likely to the path meets all of the other objectives as well.
correct the situation. Increasing the flow rate to 800 gal/rein
should be sufficient to remove cuttings effectively at angles less CONCLUSIONS
then 50, but a flow rate graatez then 1000 gal/rein would be
1. Acuttinge transport model has been presented whiohutilizes
required to remove the stationary bedsat angles greaterthan50.
fluid mechanical relationships developed for the various
Sincethemodelisastsady state solution, it cannotbeusedto
modes of particle transpork aettling,lifting, and rolling. Each
determine the circulation time needed to remove cuttings when
transport mechanism is dominant within a certain range of
they are in a moving bed. The analysis implies that one
wellbore smglee.
bottoms-up time is not sufficient, but how muchlongerthan this
2. Themodelpro~desameans ofanslyzingcuttings transportas
is needed to remove all cuttings is unknown. Cutfmgs in a
a function of operating conditions (flow rate, penetration
stationary bed cannot be removed by circulation alone unless the
rate), mud properties (denei@, rheology), well configuration
mitical flowrate isexceeded. Suchbedscenoften beremovedonly
(angle, hole size, pipe size), and cuttinge properties (density
by mechanical action via pipe rotation and e.xiaImovement. The
size, angle of repose, bed porosity).
work of Raei15indicates that a stationary bed can be tolwated if
3. Model predictions zwein good agreement with experimental
the cross-sectional areas of the bottomhole assembly and bit are
cuttings transport data for flowratesbelowcritical conditions.
lees then the area available for flow. For the exemple in Table 3,
Predicted flow rates for cxitica.1 transport, i.e., no bed
this ereais 66.8 in.2, 68% of the open-hole annuhrareaat 6310 ft.
formation, are lower then those determined visually in flow
Extended-Reach Drilling. The world record extended-reach loop experiments.
wells drilled by Statoil in 19912s and 1992/9329have been wall 4. This versatile model. in ita PC format. has been used to
documented. considerable hole cleaning-related problems were examine several situations where poorcuttinge trsnsporthad
experienced when drilling the 17Yz-in.interval on the C3 well in bean reeponsl%lefordrillingproblems. Themodelisuseful for
1991. Thisintervalwae drilled from5220 fttoaftidepthof9460 assessing the problems caused, for identif~ng potential
ftfollowingonesidetrack. Theholeanglesrangedfrom 60 to71. solutions, and for designing well paths for optimal hole
Based on this experience, the 17yz-kI. interval on the next cleaning.
extended-reach well, the C2, was planned and drilled with lower
148
.
SPE 28306 R.K. CLARKANDK. L. BICKHAM 11
149
.
10. Ford, J.l!, et al.: Experimental Investigation of Dr_~ed 25. Milne-Thomson, L.M.: Theoretical Hydrodynamics, 4thad.,
Cuttings Transport in Inclined Borehole, paper SPE 20421 The Macmillan Compsmy New York (1960) 404405.
presented at the 1990 SPE Annual Technical Conference end
26. Wallis, G.B. end Dobson, J.E.: The Onset of Slugging in
Exhibition, New Orleans, Sept. 2326.
Horizontal Stratitied Air-Water Flow, Intl. J. Mzdtiphaae
11. Siffermen, T.R.sndBecker,T!E.: HoieClecminginFull-Scale Wow (1973) 1,173-193.
Incliied Wellbore, SPEDE (June 1992) 115 120.
27. M, h; The Behavior of Drilled Cuttings in Washout
12. Luo, Y. end Bern, F!A.: Flow-Rate Predlctione for Cleaning Sections, MS thesis, U. Ms% Tulsa, OK (1979).
Deviated Wells, paper IADC/SPE 23884 presented at the
28. Njaerheim, A. and Tjoettq H.: New World Record in
1992 IADC/SPE Drilling Conference, New Orleans,
Extended-Reach Drilling From Platform Statfjord C,
Feb. 1821.
paper IADC/SPE 23349 presented at the 1992 L4DC/SPE
13. Ford, J., et al.: Development of Mathematical Models Drilling Conference, New Orleans, Feb. 1821.
Describing Drilled Cuttinge Transport in Deviated Wells,
paper 93-1102 presented at the 1993 CADE/CAODC Spring 29. Alfsen, T.E., et al.: Pushing the Limits for Extended Reach
Drilling Conference, Calgary Apr. 14-16. Drilling: New World Record tlom Platform Stut~ord C,
WellC2, paper SPE26350 presentedat the 1993 SPEAnnual
14. Lcitseri,T.I.,Pilehvari,A.A., and~ar, J.J.: Development
. . ofa Technical Conference and Exhibition, Houston, Oct. 3-6.
New Cuttin@ Trensport Model for High-hgle Wellbores
Including Horizontal Wells, paper SPE 25872 presented at 30.Hill, R.H,: Tha Mothemutiad Tkeo~ of Plasticity,
the 1993 SPE Racky Moumtein Regiourd/Low Permeability 1986 Reprint, Oxford Urdvereity Press, New Yorlq (1950)
Reservoirs Symposium, Denver, Apr. 12 14. 128-160.
15. Rasi, M.: Hole Cleaning in Large, High-Angle Wellbores~ 31. Perry, R.H, and Chilton, C.H.: Ckemical Engineers
paper IADC/SPE 27484 presented at the 1994 IADC/SPE Handbook, 5thcd., McGraw-Hill Book Company New York,
Drilling Conference, Drdlaa,Feb. 1518. NY (1973).
16. Zemor% M. ,md Hanson, F!: Rules of Thumb to Improve 32: Beris, A.N., et sd.: dreeping Motion of a Sphere Though a
High-AngleHole Cleaning,Pet. Eng, Intl. (Jan. 1991)4446, Bingham Plastic, J Fluid Mesh. (1985) 158, 219244.
48,51.
33. Zemora, M. end Bleier, R.: Prediction of Drilling Mud
17. Zamora,M, end Henson, F!:MoreRulesofThumb to Improve Rheology Using a SimpM,ed Herechel-Bulkley Model, J.
High-Angle Hole CIeaning,Pet. Eng. Zntl.(Feb. 1891) 22,M, PressureVesselTech,, Trans. ASME, (Aug. 1977)88,485-490.
2627.
34. Seffmen, PG.: The Lift on a Small Sphere in a Slow Shear
18. Einstein, H.A. rindEl=Samni, E.A.: Hydrodynamic Forces Flow, J, Fluid Mechanics, (1965) 22, Part 2,385-400.
onaRough Wrdl,Reviews ofModernPhysics (1949) 21, No. 3,
35. SefTman,PG.: Corrigendum~J. FluidMechanics, (1968) 31,
520524.
Pert 3,62%
19. E1Semni, E.A.: Hydrodynamic Forces Acting on Particles
in the Surface of a W,resm Bed, PhD disseti-ation, 36. UMherr, I?H.T., Le, T.N., rmd Tiu, C.: Characterization of
U. California, Berkeley, CA (1949). InelasticPower-Law Fluids Using Falling Sphere Da@ The
Canudian J. Chem. Eng. (Dec. 1976) 54, 497502,
20.Coleman, N.L.: A Theoretical and Experimented Study of
Drag and Lift Forces Acting on a Sphere Resting on a 37. Clii, R., Grac+ J.R., sad Weber, M.E.: Bubbles, Drops, and
Hypothetical Stresmbed, Proceedings 12th Congress of the Particles, Academic Pres~ New York (19781. -
InterrsationalAssociatwn forHydraulicResearch, FotiCo1lins 38. Beyer, WH., cd.: CRC Standard Mat~emuticol Tables, 25th
(1967) 3,185-195. Edition, CRC Press Inc., W Palm Beach, FIorida, (1978) 143.
21.Chepil, W?S.: The Use of Evenly Spaced Hemispheres to 39. Benedict, R.F!: Fundamentals of Pipe Flow, John Wiley &
Evaluate Aerodynamic Forces on the Soil Surface, Trans., Sons, New York (1980).
American Geophysical Union (1958) 39, No. 3,397-404.
40.Dodge, D.W?and Metsner, A.B.: Turbulent Flow of Non-
22. Wicks, M.: Transport of Solids at Low Concentration in
Newtonian Systems, A.I.Ch.E. J. (1959) 5, No. 2, 189204.
Horizontal Pipe, iddvances in SolidLiquidFlaw in Pipes
andItsApplicotwn, I. Zandi (cd.), PergamonPress, New York. 41. Dodge, D.WandMetzner,AB.: Errat%A,LCh.~iJ (1962)8,
(1967) 101-124. No. 1,143.
23. Davies, T.R.H. end Samad, M.WA.~Fluid DynamicLift on a 42. Govier, G.W and Asiz, K.: Tke Flow of Complex Mixtures in
ParticlqJ. HydraulicsDiv&ion, ASCE, (1978) 104,No. HY8, Pipes, van Nostrand Reinhold Compeny, New York (1972).
11711182. -
43. Torranca, B.McK.: Friction Factors for Turbulent
24. Blevins, R.D.: Ap~lied Fluid Dvnumics Handbook. Van Non-Newtonian Fluid Flow in Circular Pipes, Tks South
Nostr~d Reinhofi-Company, Ne{York (1984) &can Mech. Eng. (1963) 13, No. 3, 8991.
150
,.. .
,
SPE 28306 R. K. CLARKANDK, L, BICKHMI 13
Force Due To Pressure Gra&ent. The differential force z = ~Y+ khyn (A-12)
actimgin the z-direction due to a pressure gradient is where ~ is the yield streee, kh is the consistency index,
dFAp = (Pl P2) ~COS2 ~df3 (A-2) t= dtidr is the shear m~ (YsO), and n is the behavior index.
(When T s ~ y = Oend the strtis are equal to zero. In other
where the upstream and downstream pressure difference can be words, the plugs interior behaves es if it were an inelastic solid
eapressed as moving at a velocity of UP)
PI -Pz = rd sin~. (A-3)
PI is the upstreein pressure, PZ is the downstream presswe, d ie Lift and Drag Coefficient Models. Saffman developed an
the diameter of the sphere, 13isenenglemeaeured fromthex-axis, emdyticel model of the lateral forces acting on a sphere in a
and Iis defined in Equation (11). The preesure force can be found uniform shear flow in a Newtonian fluid.w~35Saffinens theory is
by integrating Equation (A-2) from Otcin/2. The result ie applied to tie ~ttinge trensportbyutinga%ynolds number that
Fm = Ilcds/6 (A-4) is based on the apparent viscosity of the mud surrounding the
cutting; namely,
SettlingVedocity CorrectiouFactcrs. Perry and Chiltongive
aprocedureforcrdculatingthehinderedsettlingeffect @q. 5-224, R% = QdU/~, (A-13)
p. 5-64),s1 They present agraphical method (Fig. 5-82, p. 5-65) for where
determining the exponent, n in Equation (A-5), as a function of
KB = ~Y/YP + %yp- l). (A-14)
Rep Equations (A-6), (A-7), end (A-8) were chosen to fit their
s-shaped curve within 370error. UMherretei. present amethodto celeulatetheaverage sheerrate
of a power law fluid flowing past a sphere graphically, The
u. = F,[c>R%, U:,] =_ UA (1 - C)n (A-5) following is a fit of their r.e801k3e
where
n =. e0.0811y
Sgn(x)
-1.19
(A-6) p= %[+-351 A-15)
where U is the velocity of the fluid relative to the particle. If the
y = (0.0001 + 0.865 1X1-9V3 - A-v
particle is stationary, the velocity is the axial velocity ahove the
and .. .,, cuttings bed at a point that would be occupied by the cuttings
x = -1.24 hl(ReP) 4.59. (&8) center if it were in place.
A correction for the settling velocity of the, en~elop: that E1Samnilg end Einstein and E1Sm@ls present results
eurro~de a cutting se~tling in: mud tith a yield stress cm be of the dynamic forces due to a flowing stream acting on rocks
estimated se follows. The settling velocity of the particle and protmding above a sediment bed. Their studies focused on a
envelope system can be found horn the continuity equatio~ turbulent-water stremn flowing over abed of rocks. This end the
namely, Saffman models are combined as follows
. .
.
151
.-
.
582[~~cLs2cm
B = l-~,
152
.,
SPE 28306 R.K.CLARK AND K.L.BICKHAM 15
Table 3
Survey Meas. Hole Hole Pipe Mud ECD Pressure cuttings Flow Bed
Point Depth Ang. Diem. OD Vel. Circ. Total Circ. TotiJ lwea Ht.
(ft) (deg) (ii.) (ii.) (fpm) (Ppg) (psi) (psi) % % % (in.)
*Cuttinge bed may avelanche when circulation stops if hole angle is less than 50 degrees.
I 153
.
.,:. .,
. ..
-.
_.