Professional Documents
Culture Documents
Summary the SCFF while evaluating the effect of the following variables:
This paper presents a new design model that will enable the drilling flow rate, angle of inclination, mud rheology, mud density, cuttings
engineer to select the proper hydraulics for problem-free drilling in size, drillpipe eccentricity, and rate of penetration (ROP).
high-angle holes (from 55 to 90° from vertical). Empirical correla- The model presented in this paper was developed based on the
tions have been developed after carrying out an extensive experimen- data gathered at angles of inclination ranging from 55 to 90° from
vertical. Table 1 shows the rheological properties for the three un-
tal study of cuttings transport in a 5-in. full-scale flowloop. The model
weighted and two weighted water-based muds that were tested in
predicts the required critical transport fluid velocity (CTFV), the av-
addition to water.
erage cuttings travel velocity (CTV), and the annular cuttings con-
The three cuttings sizes, by mean weight distribution, were 0.275
centration under most given sets of drilling operating conditions. in. (large), 0.175 in. (medium), and 0.09 in. (small), and the cuttings-
bed porosities (f) were 41, 36, and 39% respectively.
Introduction The CTFV and the SCFF were found experimentally using a posi-
The majority of the previous studies on cuttings transport have been tive (+62%) drillpipe eccentricity, corresponding to the pipe resting
qualitative examinations of the effects of different variables.1-6 A on the tool joints 0.5 in. above the low side of the annulus. This was
few attempts to model the rather complex nature of cuttings trans- done to simulate the worst condition in terms of cuttings transport.
port have been made.6,7 This study focused on combining the exper- Three cuttings injection rates of 10, 20, and 30 lbm/min, which
imental results with basic theoretical principles to develop an empir- correspond to a ROP of 27, 54, and 81 ft/hr in a 5.0-in. hole, were
ical predictive model for cuttings transport. investigated. The results showed that the effect of rev/min was neg-
First, an extensive experimental test program investigated all ligible for the experimental setup used in this study (drillpipe lim-
variables believed to control annular hole cleaning for angles of in- ited to only rotate around it’s own axis by centralizers). The pipe was
clination from 55 to 90°. The experimental part of this study focused rotated at 50 rev/min throughout the experiments.
The effect of the above variables will be presented when compar-
on the annular fluid velocity needed to prevent cuttings from depos-
ing the model predictions and the experimental results in the “Re-
iting in the wellbore, a concept that has been used in horizontal pipe
sults and Discussion” section of this paper.
flow in slurry transport. In addition, lower fluid velocities such that
a cuttings bed would form in the annulus were also investigated.
Development of Equations To Predict the CTFV
Secondly, an empirical model based on the experimental results
was developed to predict the minimum fluid velocity needed to keep The CTFV is found by adding the average CTV to the equivalent
all cuttings moving, the average CTV, and the cuttings concentra- slip velocity (ESV), as shown in Eq. 17. The ESV is defined as the
tion in the annulus for any fluid velocity lower than the minimum. velocity difference between the cuttings and the drilling fluid.
Subcritical Fluid Flow (SCFF). If the annular fluid velocity is low- Q inj
er than the CTFV, cuttings will start to accumulate in the wellbore. and V cut + ; . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
A annC conc–fr
Any flow rate corresponding to an annular velocity below the CTFV
is referred to as SCFF. converting from Qinj (volumetric injection rate) to ROP, as in
Experimental Setup ǒ Ǔ
hr
ft 3
R p ft + Q inj sec ǒ Ǔǒ3, 600hr secǓǒA 1
hole ft 2
Ǔ; . . . . . . . . . (4)
The experimental testing was conducted in a 5-in. diameter annulus,
which was 35 ft long with a 2.375-in. rotating inner (drill) pipe. The substituting for Qinj , from Eq. 4, into Eq. 3, while expressing con-
drillpipe eccentricity varied from negative (*62%) to positive centration in percent, as in
(+62%). The cuttings were injected into an annulus through an au-
Rp
ger system, while the fluid was pumped from a mud tank. After exit- V cut +
ƪ ǒ Ǔƫ
. . . . . . . . . . . . . . . . . . . . . (5)
ing the annulus, all cuttings were weighed on a scale. A pipe
36 1 * C conc
A hole
Summary of Experimental Work
The experimental results used to develop the model were part of a
Rp
study of more than 700 tests designed to investigate the CTFV and or V cut +
ƪ ǒ Ǔ ƫC
. . . . . . . . . . . . . . . . . . . (6)
2
D pipe
*Previously with U. of Tulsa 36 1 * conc
D hole
Copyright 1997 Society of Petroleum Engineers
Original SPE manuscript received for review 26 July 1993. Paper peer approved 12 March
1997. Paper (SPE 25872) first presented at the 1993 SPE Rocky Mountain Regional Meet-
It was found from the experimental data that the annular cuttings
ing/Low Permeability Reservoirs Symposium held in Denver, Colorado, 26–28 April. concentration, by volume, at CTFV can be expressed as a function
of the ROP for angles of inclination from 55 to 90°. Fig. 1 shows the where the apparent viscosity is given by8
cuttings concentration vs. ROP for water, three bentonite muds
(muds 1, 2, and 3), two weighted muds (muds 4 and 5), and three 5 Y p(D hole * D pipe)
different cuttings sizes. Fig. 1 can be expressed in terms of cuttings ma + mp ) . . . . . . . . . . . . . . . . . . (11)
V crit
concentration and ROP by the equation
To generalize these equations, correction factors for angle of in-
C conc + 0.01778 R p ) 0.505. . . . . . . . . . . . . . . . . . . . . . . (7) clination, cuttings size, and mud weight have been introduced. Note
The CTV can, by combining Eqs. 6 and 7, be expressed as that the Vcrit (CTFV) used in Eq. 11 is the velocity we are looking
for and is not yet known. Therefore, an iterative procedure will be
V cut + 1 required. Vcrit needs to be estimated initially. (Refer to Appendix A
ƪ ǒ Ǔ ƫ(0.64 ) 18.16
. . . . . . . . . . . . (8)
2 for a sample calculation.)
D pipe
R )
1* Angle of Inclination Correction Factor. The angle of inclination
D hole p correction factor was found by dividing the experimental CTFV
mean for the individual angles (90, 75, 65, and 55°) by the average
Eq. 8 shows that the CTV at a flow rate corresponding to the CTFV of all angles, as shown in Fig. 3 (which shows that angles ranging
is independent of mud rheology, mud weight, and angle of inclination from 65 to 80° are slightly harder to clean) or the equation
between 55 and 90°. This simply confirms that the experimental tests
were performed according to the definition of the CTFV (minimum C ang + 0.0342q ang–0.000233q 2 ang–0.213. . . . . . . . . . . . . (12)
fluid velocity required to maintain cuttings movement). Note that if
the average cutting is not traveling at this velocity, the slowest-mov- Cuttings-Size Correction Factor. The cuttings-size correction
ing cuttings will start to accumulate in the annulus. factor, shown in Fig. 4, is generated by dividing the average results
of large, medium, and small cuttings by that of the large cuttings,
ESV and Its Correction Factors. The ESV can be predicted by cal- while using mud 1. It is also represented by the equation
culating the apparent viscosity (from Eq. 11), as shown in Fig. 2.
The figure was generated by using Eq. 8, Eq. 17, and the experimen- C size + –1.04D 50 cut ) 1.286. . . . . . . . . . . . . . . . . . . . . . . (13)
tal results of water and muds 1, 2, and 3 with the large cuttings size.
Using linear regression, Fig. 2 can be represented by Mud-Weight Correction Factor. The correction for mud weight
has been based on muds 2 (8.65 lbm/gal) and 4 (11.0 lbm/gal). The
V slip + 0.00516 m a ) 3.006 for m a t 53 cp, . . . . . . . . (9) test results for mud 5 (15.0 lbm/gal) were not incorporated into this
correction factor since the plastic viscosity (mp ) could not be kept at
and V slip + 0.02554ǒ m a–53 Ǔ ) 3.28 for m a u 53 cp, . . . (10) 14 cp, as the case was for the two other muds, but rather rose to 28
Fig. 1—Cuttings concentration vs. ROP at CTFV. Fig. 2—Equivalent slip velocity vs. apparent viscosity.
cp. Thus, the effect of density was not totally isolated. Fig. 5 can be where Vopen is the average fluid velocity above cuttings bed at
used to find the correction factor for mud weight, as can the equations steady state, and Vcrit is the CTFV (no cuttings accumulation).
In terms of flow rate and the corresponding area open to flow, the
C mw t + 1–0.0333(g m–8.7) g m u 8.7 . . . . . . . . . . . . . . . (14) above equation becomes
Q pump Q
and C mw t + 1.0 g m t 8.7. . . . . . . . . . . . . . . . . . . . . . . . (15) + crit , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (19)
A open A ann
Therefore the generalized ESV is expressed as where Qpump is the pump, or operating, flow rate; Qcrit is the flow
rate corresponding to CTFV; Aopen is the area open to flow above the
V slip + V slip C ang C size C mwt . . . . . . . . . . . . . . . . . . . . . . . . (16)
cuttings bed; and Aann is the total area of the annulus.
The area occupied by the cuttings bed (Abed ) can be calculated by
Determining CTFV. The CTFV can be found by adding the CTV combining Eq. 19 with the equation
(from Eq. 8) to the generalized ESV (from Eq. 16), as in A bed + A ann–A open . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (20)
V crit + V cut ) V slip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (17) to achieve
ǒ Ǔ
self, the assumption can be made that Q pump
A cut + A ann 1– (1–f) . . . . . . . . . . . . . . . . . . . . . . . (23)
V open + V crit , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (18) Q crit
ǒ
C bconc + 100 1–
Q pump
Q crit
Ǔ
ǒ1–fǓ . . . . . . . . . . . . . . . . . . . . . (24)
mud 3 and water, according to viscosity. Low-viscosity muds, or
water, perform better in high-angle wells because the fluid velocity
below the drillpipe is higher. As stated earlier, a positive eccentric
Eq. 24 is the basis for cuttings concentration calculations for any drillpipe resting on tool joints on the low side of hole will cause
given fluid flow rate that corresponds to a velocity lower then the higher-viscosity muds to divert more flow from the narrow section
CTFV. A more accurate result can be obtained by incorporating a below the drillpipe where the cuttings are transported to the open
correction factor. section above the pipe, as shown in Fig. 7.
A change in ROP and a variation in cuttings size were predicted
Correction Factor for Cuttings Concentration for SCFF. The by the model, as respectively shown in Figs. 10 and 11. To reach
prediction (calculated by Eq. 24) of cuttings concentration at a CTFV a higher velocity is needed for an increase in ROP, as shown
SCFF rate of 110 gpm is fairly close to the experimental values for in Fig. 10. Fig. 11 shows that a smaller cuttings size requires a larger
water and mud 1 data, as shown in Fig. 6. A slight overprediction flow rate to reach CTFV.
can be seen. Note that the predicted cuttings concentration for this An increase in mud weight will improve cuttings transport, as
figure is based on the actual (experimental) CTFV. shown in Fig. 12, where mud 2 (8.65 lbm/gal) is compared to mud
For higher-viscosity muds an even larger overprediction of the 4 (11.0 lbm/gal). The intention was to maintain the mp at 14 cp and
cuttings concentration was experienced. The location of the drill- the yield point (Yp ) at 14 lbf/100 ft2 while increasing the mud
pipe during CTFV testing caused the overprediction of the cuttings weight. This was not successful for mud 5 (15.0 lbm/gal) where the
concentration. As Fig. 7 shows, a positive eccentricity requires a mp increased to about 28 cp. This higher viscosity, for the positive
higher flow rate to reach CTFV because the mud velocity slows eccentric annulus, affected the enhancement gained by a higher mud
down in the narrow section of the annulus, causing a diversion of density. For this mud the density effect could not be isolated.
flow to the more open area. The narrow section of the annulus is lo-
cated below the pipe, where the cuttings are transported in highly SCFF. The predicted SCFF cuttings-concentration values, which
deviated wellbores, for a positive eccentric pipe. This narrow sec- are based on the calculated CTFV values, are compared to the exper-
imental results in Figs. 13 and 14.
tion will be covered by accumulated cuttings for SCFF testing, and
The effect of rheology on cuttings concentration for two different
the velocity distribution above the cuttings bed will be more uni-
SCFF rates [150 gal/min (3.17 ft/sec) and 110 gal/min (2.32 ft/sec)],
form. It was found that the diversion of flow increases as the viscos-
ity of the mud increases, while the diversion effect is reduced for
lower-viscosity fluids, such as water. To compensate for this diver-
sion of flow to the more open area of the annulus, a correction factor
has been introduced for the SCFF predictions. This factor has been
generated from the mean SCFF results using all variables, as shown
in Fig. 8. In order not to underpredict the cuttings concentration, the
correction factor was estimated conservatively.
Using linear regression, Fig. 8 can be reproduced by
C bed + 0.97–0.00231m a . . . . . . . . . . . . . . . . . . . . . . . . . (25)
The generalized cuttings concentration can be found from
C bconc + C bconc C bed . . . . . . . . . . . . . . . . . . . . . . . . . . (26)
Fig. 12—Effect of mud weight for predicted vs. experimental Fig. 13—Effect of rheology for predicted vs. experimental SCFF
CTFV. corresponding to 3.17 ft/sec.
are shown in Figs. 13 and 14. Annular cuttings concentrations of mud Aopen + area open to flow above the cuttings bed, ft2
1 and mud 3 are predicted fairly closely, while that of mud 2 is over- Apipe + area of drillpipe or drill collars, ft2
predicted by about 14% for both 150 and 110 gal/min flow rates. Cang + correction factor for angle, nondimensional
By comparing Figs. 13 and 14, it can be noted that dropping the Cbed + correction factor for cuttings concentration,
flow rate from 150 gal/min to 110 gal/min caused the annular cut- nondimensional
tings concentration to increase by a factor of about 1.6. Cconc + cuttings concentration by volume at CTFV, %
Cconc-fr + fractional cuttings concentration by volume at CTFV,
Conclusion nondimensional
Based on an extensive experimental testing program, gathered in a C+ correlation factor for the ESV, nondimensional
5-in. diameter full-scale flowloop, simple empirical correlations Cbconc + cuttings concentration for a stationary bed (by
have been developed for predictions of CTV, CTFV, and the cuttings volume), corrected for viscosity, %
concentration for SCFF rates. The predictions can be read easily Cbconc-fr + fractional cuttings concentration for a stationary bed
from the charts, hand calculated, or programmed on a personal com- (by volume), corrected for viscosity,
puter. A sample calculation can be found in the Appendix. nondimensional
Please note that as a continuation of this project a hole-size correc- Cbconc + cuttings concentration for a stationary bed (by
tion factor for the CTFV predictions has recently been developed by volume), not corrected for viscosity, %
Jalukar et al.10 No correction factor was needed for SCFF predictions. Cmwt + correction factor for mud weight, nondimensional
Csize + correction factor for cuttings size, nondimensional
Nomenclature D50sut + mean cuttings size, in.
Aann + area of annulus, ft2 Dhole + diameter of drilled hole, in.
Abed + area of cuttings bed, ft2 Dpipe + diameter of drillpipe, in.
Acut + area occupied by cuttings in the annulus, ft2 Mgd + mass generated by drill bit, lbm
Ahole + area of the drilled hole, ft2 Mtm + mass transported by mud, lbm
Eq. A-10 yields 3.15 ft/sec, compared to Eq. A-6, which yielded
3.14 ft/sec. This is close enough. In other words, the ESV is not very Correction for Cuttings Concentration for SCFF. The correction
sensitive to Vcrit . Note that, for simplicity, the uncorrected Vcrit has factor can be found from Eq. 25 by using the apparent viscosity pre-
been used to calculate the apparent viscosity. viously from Eq. A-9 to give
Angle of Inclination Correction Factor. Using Eq. 12 for C bed + 0.97–0.00231 27.6 cp + 0.906. . . . . . . . . (A–21)
C ang + 0.0342 65–0.000233 65 2–0.213 + 1.026. Corrected Cuttings Concentration. From Eq. 26,
C bconc + 33.5 0.906 + 30.4%. . . . . . . . . . . . . . . . (A–22)
. . . . . . . . . . . . . . . . . . . . (A-11)
SI Metric Conversion Factors
Cuttings-Size Correction Factor. Using Eq. 13 for cp 1.0* E*03 +Pa@s
ft 3.048* E*01 +m
C size + –1.04 0.175 ) 1.286 + 1.10. . . . . . . . . . . (A-12) ft2 9.290 304* E*02 +m2
ft3 2.831 685 E*02 +m3
gal 3.785 412 E*03 +m3
Mud-Weight Correction Factor. It is not necessary to correct for
in. 2.54* E)00 +cm
mud weight because the mud weight is below 8.70 ppg. The correc- lbf 4.448 222 E)00 +N
tion factors can be found from Eqs. 14 and 15, for lbm 4.535 924 E*01 +kg
C mwt + 1.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-13) *Conversion factor is exact. SPEDC
ESV. The corrected ESV is found by combining all the correction T.I. Larsen has worked for Unocal Corp. in Lafayette, Lousiana, for
factors, as shown in Eq. 16, for the past 7 years. He holds BS and MS degrees (1988 and 1990, reĆ
spectively) from the U. of Tulsa, Oklahoma. He has conducted exĆ
V slip + 3.15 1.026 1.10 1.0 . . . . . . . . . . . . . . . (A-14) tensive testing on cuttings transport in deviated holes at the Tulsa
U. Drilling Research Project's flowloop. Ali Pilehvari is an associate
and V slip + 3.56 ftń sec. . . . . . . . . . . . . . . . . . . . . . . . . . (A-15) professor in the Dept. of Chemical and Natural Gas Engineering
at Texas A&M U. in Kingsville, Texas. He was formerly a visiting assisĆ
tant professor in the petroleum engineering department at the U.
CTFV. Add the CTV to the ESV for CTFV, as shown in Eq. 17, for of Tulsa, and he was the assistant director of the Tulsa U. Drilling ReĆ
search Projects. Pilehvari earned a BS degree in chemical engiĆ
V crit + 1.32 ftń sec ) 3.56 ftń sec . . . . . . . . . . . . . . . (A-16) neering from Tehran Polytechnic in 1972, and a PhD degree in
chemical engineering from the U. of Tulsa in 1984. J.J. Azar is a proĆ
fessor of petroleum engineering and past director of the U. of Tulsa
and V crit + 4.88 ftń sec. . . . . . . . . . . . . . . . . . . . . . . . . . (A-17) Drilling Research Project. He has extensive experience in applied
industrial drilling research and teaching. He lectures and consults
worldwide and is the author or coauthor of books in drilling engiĆ
Cuttings Concentration Prediction for SCFF neering and structures. Azar holds a PhD degree in mechanical
engineering from the U. of Oklahoma. He served as a member of
The CTFV needs to be converted to flow rate to compare it to the
the Career Guidance Committee Educational/Professional TechĆ
pump operating flow rate. nical Committee and was a student chapter faculty sponsor.
ǒ ft
Q crit + 4.88 sec Ǔǒ601 min Ǔǒ 1 ft Ǔǒ0.10559
sec 7.48 gal
3
ft 2Ǔ ,