SPE 81746

Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in
Horizontal and Deviated Wells in Coiled-Tubing Drilling
V. C. Kelessidis, SPE, G. E. Mpandelis, Technical University of Crete, Greece
Copyright 2003, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE/ICoTA Coiled Tubing Conference held in
Houston, Texas, U.S.A., 89 April 2003.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836 U.S.A., fax 01-972-952-9435.

Abstract
Coiled Tubing Drilling, grown significantly in recent years, is
normally associated with high angle to horizontal and
extended reach wells. It is, however, in these applications that
hole problems become more troublesome because of
inefficient cuttings removal. Among the many parameters
affecting efficient cuttings transport in Coiled Tubing Drilling
are pump rates, well dimensions, fluid properties, solids sizes,
solids loading and hole inclination. Several attempts have been
made to determine the optimum operating range of these
parameters but complete and satisfactory models have yet to
be developed.
The purpose of this paper is to provide a critical review of the
state of the art on efficient cuttings transport during Coiled
Tubing Drilling, present the critical parameters involved,
establish their range according to what is observed in practice
and propose a different approach for predicting the minimum
suspension velocity. Finally the laboratory system that has
already been set up is presented. Its primary purpose is to
allow the gathering of good quality data, missing from the
literature, which could enhance our understanding of the flow
of solid liquid mixtures in annuli.
Introduction
The advantages of Coiled Tubing Drilling (CTD) are
numerous and have been indicated and proved in practice by a
large number of investigators. A significant drawback is the
difficulty for efficient cuttings transport primarily because the
pipe is not rotated.
Cuttings transport during drilling (either conventionally or
with Coiled Tubing) has a major impact on the economics of
the drilling process. Inefficient hole cleaning from the cuttings
can lead to numerous problems such as stuck pipe, reduced

weight on bit leading to reduced rate of penetration (ROP),

transient hole blockage leading to lost circulation conditions,
extra pipe wear, extra cost due to additives in the drilling fluid
and wasted time by wiper tripping.
These many problems have prompted significant research into
cuttings transport during the past 50 years. Excellent reviews
on the subject have been given in the past1-3. Pilehvari et al.1-2
state that fluid velocities should be maximized to achieve
turbulent flow and mud rheology should be optimized to
enhance turbulence in inclined / horizontal sections of the
wellbore. Turbulent flow of non-Newtonian fluids needs much
more work and should be extended to include pipe rotation
and dynamics for conventional drilling. Future work should
focus on getting more experimental data, validation of fluid
models, cuttings transport mechanistic models verified by
comprehensive experimental data. Azar & Sanchez3 conclude
that a combination of appropriate theoretical analyses
(complete free body diagrams, accurate rheological models,
accurate annular flow models), experimental studies (extensive
testing concentrating on individual variables or phenomena),
statistical modeling (rheological models, unstable cuttings
transport conditions), and high tech research facilities
(accurate measurement of pertinent variables, analysis of
video to develop flow pattern maps) will be necessary for
further progress.
While many cuttings transport problems were addressed quite
successfully for conventional drilling in vertical, inclined and
horizontal wells in the past, the increase in activity of CTD has
called for renewed interest into cuttings transport problems in
horizontal and highly inclined annular geometries with no
rotation of the inner pipe.
In recent years there have been several theoretical, semitheoretical and experimental investigations for assessing the
important parameters for efficient cuttings transport in highly
inclined and horizontal geometries during CTD4-10 or
conventional drilling but not taking into account the rotation of
the inner pipe11-15.
Despite these efforts, there is still lack of good quality
published data against which models can be compared with.
Many models use data for validation not applicable to the
situation at hand (e.g. using data with inner pipe rotation4,12) or
even no data at all13, or compare their results with results of

V. C. Kelessidis, G. E. Mpandelis

other models11. Some authors present the data together with

their own theoretical analysis6-8 but they give very limited
information on experimental parameters (e.g. rheology of
fluids, cuttings concentrations, etc.).
The approach taken by many investigators on modeling
cuttings transport for highly inclined and horizontal annuli
(with no inner pipe rotation) is that of two or three layer
model. The basic model proposed for annuli4,6,9,11-15 is adopted
from the one proposed for solids transport in pipes16, later
extended by the same authors17-19.
The steady state models are based on mass balance equations
for solids and liquid plus momentum balance equations for the
two or three layers resulting in a system of coupled algebraic
equations. Closure relationships that describe the interaction
of the two phases are needed in order to solve these equations
and these are taken from published correlations.
Flow Patterns
During the flow of solid - liquid mixtures in horizontal
conduits, the liquid and solid phases may distribute in a
number of geometrical configurations depending on flow
rates, conduit shape and size, fluid and solid properties and
inclination. Natural groupings, or flow patterns, exist within
which the basic characteristics of the two-phase mixture
remain the same.
The main parameters determining the distribution of solids in
the liquid, i.e. the flow patterns, are the liquid velocity, the
solids loading and the properties of liquid and solids (rheology
and density of liquid, density, diameter & sphericity of solids).
Experimental observations of solid liquid flow in horizontal
pipes and annuli, even at low solids concentrations, suggest
the following flow patterns (depicted in Figure 1), in the
direction of decreasing flow rate (or velocity)4,7,12,16,19-21.

SPE 81746

been given different names like, limit deposit velocity,

suspension velocity, critical velocity (Fig. 1c).
Further reduction in liquid velocity results in more and more
solids deposited resulting in three layers (Fig. 1d). A bed of
solids that is not moving, forming a stationary bed, a moving
bed of solids on top of the stationary bed and a heterogeneous
liquid solid mixture above. There is a strong interaction
between the heterogeneous solid liquid mixture and the
moving bed, with solids deposited on the bed and re-entrained
in the heterogeneous solid liquid mixture. There is a point of
equilibrium where, with the increase in height of the solids
bed, the available area for flow of the heterogeneous mixture
is decreased resulting in higher mixture velocities and hence
an increase in the erosion of the bed by the mixture.
At even lower liquid velocities the solids pile up in the pipe
(or annulus) and full blockage may occur. Experimental
evidence and theoretical analysis indicate that this may occur
at relatively high solids concentration, not encountered during
normal drilling operations. It may occur, however, if cuttings
transport is inefficient resulting in high solids concentration,
especially in sections where large cross sectional areas exist
(e.g. in annulus washouts).
It should be stressed that the solids concentration in the liquid
is fairly low during normal drilling operations, and for CTD
rarely exceeds 2 4% by volume (Appendix A).
Two Layer Modeling
The two layer model (moving bed & clear liquid in laminar
flow) proposed by Gavignet and Sobey22 made a basic
advance but data did not conform to model predictions. The
model was extended4,13,21,23 to account for suspended solids in
the liquid layer and covered the flow patterns described above.
This resulted in a system of four algebraic equations and one
integral equation for turbulent diffusion of solids.

At high liquid velocities the solids may be uniformly

distributed in the liquid and normally the correct assumption is
made that there is no slip between the two phases11,12,14,16, i.e.
the velocity of the solids is equal to the velocity of the liquid.
This flow pattern is normally observed for fairly fine solids,
less than 1mm in diameter, not normally occurring during
drilling applications. This flow pattern is called the fully
suspended symmetric flow pattern (Fig. 1a).

Following these approaches, for steady state conditions and for

horizontal concentric annulus, the two layer model is as
follows (Figure 2):

As the liquid flow rate is reduced there is a tendency for the

solids to flow near the bottom of the pipe (or outer pipe of the
annulus), but still suspended, thus creating an asymmetric
solids concentration. This is called the asymmetric flow
pattern with the solids still moving with the liquid (Fig. 1b).

Mass balance for the solids gives,

A further reduction in the liquid flow rate results in the

deposition of solid particles on the bottom of the pipe. The
solids start forming a bed, which is moving in the direction of
the flow, while there may be some non uniformly distributed
solids in the liquid layer above. This is the moving bed flow
pattern and the velocity below which this is happening has

There is a layer of liquid on the top side of the annulus

containing suspended solids and a layer of solids on the
bottom of the outer pipe which may be moving, albeit at a very
low velocity.

U s As C s + U B AB C B = U M AM C M

(1)

And mass balance for the liquid gives,

U s As (1 C s ) + U B AB (1 C B ) = U M AM (1 C M ) (2)

The momentum equations for each of the two layers are

As

dp
= s S s i Si
dz

(3)

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 3

AB

dp
= FB B S B + i Si
dz

(4)

For the mean concentration of the solids in the liquid layer,

C s , the turbulent diffusion equation is solved (Appendix B)
yielding

C
Cs = B d o2 * Into d i2 * Inti
2 As

(5)

where,
Int o =

/2

u pdo
exp
D

sin o
h

2
d
o

(cos ) 2 d (5a)
o
o

B = s =

(15)

Dhs =

4 As
S s + Si

(16)

D hB =

4 AB
SB

(17)

Martins and Santana13 also use the Fanning friction factor and
from experimental work24 they use, for turbulent flow,
0.7
f s = 0.00454 + 0.645 N Re,
gn

(18)

with
Inti =

/2

u p do
exp
D

sin i
h

(cos i ) 2 d i (5b)

2(d / d ) d
o i
o

There are 5 equations and 5 unknowns, namely, h, Us, Cs, UB,

dp/dz. For the solution we need closure relations for the shear
stresses, the friction force, particle settling velocity up and
dispersion coefficient, D (Appendix B). The shear stresses are
estimated by:
1
s = f s sU s2
2

(6)

1
i = f i s (U s U B )2
2

(7)

1
f B BU B2
2

(8)

B =

Walton4 expresses the friction factors in a standard way, but

no indication is given whether the ones he uses are for
Newtonian or non-Newtonian fluids (presumed Newtonian).
Doron et al.16 use the following relationships valid for
Newtonian fluids, for turbulent flow,
fs =

0.046
0.2
N Re,
s

(9)

and for laminar flow

fB =

16
N Re, B

(10)

with

N Re, s =

sU s Dhs
s

N Re, B =

BU B DhB
B

(11)

N Re, gn =

n
8 U s2 n D hs

2(3n + 1)
K

(19)

For the interfacial friction factor, f i , most researchers use the

expression proposed by Televantos et al.25 ,
1
2 fi

d p / D hs
2.51
= 0.86 ln
+
3.7
N Re, s 2 f i

(20)

which assumes an interfacial roughness equal to a particle

diameter, dp. Accounting for particle collisions with the bed
and for entrainement and deposition of particles, which tend to
increase f i , Televantos et al.25 use the factor (2* f i ) instead
of ( f i ) in the above expression (20), which is the Colebrook
formula for Newtonian fluids flowing in rough pipes. And as
stated26 according to the study of Televantos, the magnitude of
f i is not of critical importance in the model.
However, Martins et al.27 found that the value of f i can have a
dramatic impact on the results of the model. By measuring the
bed thickness, they correlated the interfacial friction factor
with various parameters. They got the dependence of f i on
NRe,gn, given by equation (19), the fluid behavior index, n, the
particle diameter, dp, and the hydraulic diameter of the
suspension, Dhs. Their results, however, do not show any
significant correlation with neither of these parameters, as a
simple inspection of their proposed figures easily reveal.
Nevertheless, they proceed and propose a relation such as

1.07116
f i = 0.966368 N Re,
(n )2.360211 d p / Dhs 2.34539 (21)
gn

which is also used in their next paper14.

(12)

s = p C s + (1 C s )

(13)

B = p C B + (1 C B )

(14)

FB in equation (4) is the friction force opposing the motion of

the bed of solids. At the point of slip of the bed, FB is equal in

magnitude to the maximum sliding friction force between the
particles and the wall, Fmax ,

V. C. Kelessidis, G. E. Mpandelis

S
FB = Fmax = FN = p gC B AB + i i (22)
tan

where the first term in the r.h.s. represents the submerged

weight of the bed and the second term represents the Bagnold
stresses28-29. Bagnold showed that when a fluid flows over a
deposit of solid bed, there exists a normal stress at the
interface of solids - wall associated with the shear stress
exerted by the fluid on the solids. However, caution should be
exercised, since Bagnolds experiments were conducted at
relatively high solid concentrations (> 15%20,26).
Martins and Santana13 used the basic formulation described
above. They do not reference how they estimate the solids
dispersion coefficient, D , needed in the diffusion equation
(5). They presumably use the one proposed elsewhere16,
described in Appendix B. For u p , they refer to a procedure for
nonNewtonian fluids30, but do not give the equations.
Published equations give u p for nonNewtonian31-35 and for
Newtonian liquids16,26,36. Their model predicts that increasing
the flow rate and the density of the liquid are the most
effective ways to enhance cuttings transport. Rheological
parameters have only a moderate effect, but no comparison
with experimental data is given. However, in a subsequent
paper14 they found that the type of rheological model used
influenced significantly the predictions. No suggestions,
however, is given on the appropriate rheological model,
supported by experimental data.
Walton4 also used the above formulation. For u p he used
correlations for Newtonian fluids, as Doron et al.16, while for
D he proposes a slightly different approach, also described in
Appendix B. He compares his simulation results, using water,
with data taken with inner pipe rotation at 50 RPM37, hence, as
he states, no quantitative agreement is expected. The
simulation predicts the trend of the data, however, at high flow
rates data show no bed for concentric annulus which is
different from his predictions. He produces a flow regime map
and a map that shows how the minimum flow rate for
complete suspension varies with fluid viscosity and particle
concentration. Fluids of moderate viscosities are more
efficient than low viscosity fluids such as water or high
viscosity gels, with optimum viscosity around 20 30 cp at
170 s-1. Furthermore he predicts that cleanouts in horizontal
and highly deviated wells require pump rates an order of
magnitude higher than vertical wells.
Kamp and Rivero11, who made an excellent review of layer
models, use also a two layer model with eccentric annuli. They
use mass flux of cuttings per unit interface that are deposited,
s, dep , or resuspended, s, susp , mass balance equations for
solid, liquid and mixture and momentum equations for the
heterogeneous layer and the bed. They use also the diffusion
equation for the concentration in the suspension layer, but
instead of using C B in the solution, they use the average
concentration, C s . The deposition of particles, s , dep , is taken

SPE 81746

as proportional to Cs and Us. The resuspension flux, s , susp is

taken as a function of interfacial shear stress, i . The settling
velocity, u p , is derived for Newtonian fluids through standard
procedures16,26,36, while some reference is given on the effect
of non-Newtonian fluids. There is no equation given for the
solid dispersion coefficient, D , except referencing appropriate
modeling38, also followed by Doron et al.16 as well as other
work39, where, though, no specific equation is given for the
prediction of D .
Their results, using as a base case a rather large annulus for
CTD, d 0 = 17.5in and d i = 9.0in and a flow rate
Q = 50 gpm resulting in very low velocity in the annulus
( 0.092 ft / s !) and ROP = 50 ft / hr (hence volumetric
concentration ~ 17%) showed that starting from well
distributed solids at the entrance there is almost immediately
bed formation in the annulus. The height of the bed then
remains constant along the flow channel. Similar results are
also derived for the other pertinent parameters. The bed height
decreases as the flow rate is increased but the rate of decrease
is much smaller at higher flow rates, contrary to what is
expected. Results are not very sensitive to mud viscosity, in
contrast to experimental and field evidence which show that
turbulence promotes cuttings transport. They compare their
results with predictions of correlation based models40-42.
Predictions show similar trends but they are far from being
quantitatively close. It is recognized that the model
overpredicts cuttings transport at a given liquid flow rate.
They state the need for good closure terms with respect to
particle resuspension and particle settling velocities.
Li and Walker6 presented results of their experimental study
using various empirical correlations, devised from their own
data as well as from other investigators. They assume that
cuttings erosion follows logarithmic expression with time.
They developed a computer model based on dimensional
analysis and using these correlations they study the sensitivity
of predictions on various important parameters (including
underbalanced conditions). Most important variable is the
liquid velocity, but in general, no quantitative data is given.
They extended their study7 to cover the effects of particle
diameter, fluid rheology and eccentricity. They conclude that
fine particles are easier to clean and spherical particles with dp
~ 0.03in. pose the greatest difficulty. Fluid rheology plays a
significant role with low viscosity fluids in turbulent flow
giving optimum results for hole cleaning, similar to other
results4. Furthermore, the critical velocity for full suspension
is higher for fully eccentric annulus compared to the
concentric annulus. However, the differences observed among
the operational parameters were not substantial and in addition
they tested only three particle sizes ( 0.006in,0.03in,0.275in ).
A similar two layer model has also been used for
underbalanced drilling under transient conditions31. Here the
two phase (gas liquid) mixture in the suspension layer is
considered as pseudohomogeneous with fixed properties.
The authors also consider deposition fluxes, derived from

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 5

hindered settling of solids and erosion (or resuspension) fluxes

as a function of interfacial shear velocity, similar to other
approaches11. They allowed slip between the particles and the
fluid both in the suspension layer and the bed and accounted
for the non-Newtonian behavior of the liquid. They use a
correlation for solids settling velocity for non-Newtonian
fluids but use hindered settling expression proposed for
Newtonian fluids, as elsewhere16.
Their results (run for water and mud only, without a gas
phase), showed that for the fully suspended asymmetric flow
pattern, cuttings velocity matches that of liquid, i.e. there is no
slip between the two phases. However, when a bed is formed,
then the solids in the suspension move at velocities much
lower (about 50%) of the liquid velocity. Their predictions
compare favorably with their experimental results, except in
the cases of very dilute solid concentrations (less than 0.05%
by volume) because a smooth interface (assumed in the model)
did not really existed. Effective transport is achieved when
cuttings velocity is greater than zero and they establish a range
of parameters from which they can determine the minimum
inlet cuttings concentration for this to be achieved.
Three Layer Modeling
The inadequacies in the predictions from the two layer
models when compared to limited data and the extension of
the two layer to the three layer model for flow of solid liquid
mixtures in pipes17 led to the extension of the two layer model
in annulus to the three layer model9,12.

The development follows that of the two layer model

approach, with the inclusion of a stationary bed below the
moving bed. For a horizontal concentric annulus, the three
layer models, depicted in Figure 3:
The mass balance equations are
AsU s C s + AmBU mB CmB = AM U M C M

(23)

for solids, and

AsU s (1 C s ) + AmBU mB (1 CmB ) = AM U M (1 CM )
(24)

for liquid.
For the momentum equations we have, for the suspended
layer,

As

dp
= s S s smB S smB
dz

(25)

dp
= FmBsB FmB mBsB SmBsB
dz
mB SmB + smB S smB

(26)

dp
+ FmBsB + mBsB S mBsB FsB
dz

The concentration in the suspension layer is derived from the

diffusion equation, as for the two layer model, with the
interface now being ( mB ), rather than ( sB ).
The solution to the above equations requires closure
relationships for the shear stresses and the friction forces,
which are taken through the use of Reynolds number and
friction factor relationships, as for the two layer modeling.
Nguyen and Rathman12 use the above formulation but avoid
using the diffusion equation. Instead, they predict the
thickness of the moving bed, hmB , based on analysis of
Bagnold forces, as derived by Wilson and Tse43
hmB =

si

( p ) gU mB tan

(28)

where the moving bed velocity, U mB , is predicted from

turbulent boundary layer theory as
U mB =

4 smB /
Kv
3

(29)

The authors predict full erosion of bed at annular velocity of

U M 3.8 ft / s for flow of water in a 5 by 1.9 in. annulus,
with particle diameter of 0.25 in., ROP = 50 ft / hr
corresponding to 0.42% solids volumetric concentration at this
velocity. This compares favorably with other experimental
results44-45 who measured minimum transport velocity of
3.3 ft/s for similar conditions. Various runs of the simulator
showed that liquid density and annulus eccentricity had
significant effect. The use of the assumed value of CB = 0.62
(instead of the standard used of 0.52) had minor effect, while a
low viscosity bentonite mud showed better cuttings transport
compared to that of water.
Cho et al.9 follow the above three layer formulation but use the
diffusion equation (equation 5) and an experimental
expression for the settling velocities of solids through nonNewtonian fluids46
u 2p + 0.45u p ( / d p ) exp(5 p )

(30)

with p taken as 0.8 for drill cuttings. Hindered settling is also

considered, taken from the correlation developed by Thomas47

uh = u p exp(5.9Cs )

(31)

For the particle dispersion they use the approach of Walton4.

Their base case is a low viscosity bentonite mud, with

and for the stationary bed,

AsB

The equation for the stationary bed (27) serves as a condition

to be satisfied whenever a stationary bed is predicted but it is
not part of the solution17.

19.45exp(5 p )(d p )( / ) = 0

while, for the moving bed layer,

AmB

where the various stresses and friction forces are shown in Fig.
3.

(27)

V. C. Kelessidis, G. E. Mpandelis

n = 0.7 and K = 0.605lb f s n / 100 ft 2 , in a 5 by 1.9 in.

annulus, p = 21.7 ppg, = 9.16 ppg and dp = 0.09 in.

Their results show that for liquid annulus velocities in the
range of 1 5 ft/s, typical velocities encountered in CTD
applications, a bed of cuttings is formed. At about 2 to 3 ft/s
suspension starts to occur, similar to previous results12, and the
critical velocity for full suspension is around 4 to 5 ft/s. The
most significant parameter is the annulus liquid velocity. It
should be of the order of 3.6 to 4.6 ft/s, compared to
conventional drilling of 2 to 3 ft/s because both the minimum
pressure drop and a minimum height of stationary bed is
obtained at these flow rates. There is a slight effect of
rheological parameters. When the moving bed vanishes, the
model does not reduce to a two layer model (as it should do as
a limiting case). Instead, a new model covering the two layers
is solved. No comparison is made to the predictions versus that
of two layer model to justify the added complexity of the three
layer model.
Discussion on the Layer Models
Careful analysis of the published results shows that a two or
three layer pattern will form almost immediately, even when
starting from a homogeneous distribution of solids at the
entrance of the annulus. The bottom solids concentration, or
equivalently the height of the bed, are constant at two to three
annulus hydraulic diameters from the entrance11,15!

From the proposed tentative solutions from the various

investigators it is evident that the framework for the problem
solution is similar, with main differences whether a two or a
three layer model is used and the closure relationships which
include, among others:
Solid distribution in the heterogeneous liquid - solid layer
Interfacial friction factor between the heterogeneous
liquid layer and the moving bed of solids
Whether Bagnold stresses are taken into account
Terminal velocity of solids in Newtonian or nonNewtonian fluids, taking into account the effect of
hindered settling and the effect of walls (normally not
taken into account)
Fluid friction factors for fluid and walls of annulus,
normally taken for Newtonian fluids and using
correlations developed for pipe flow.
Full suspension, derived from full erosion of beds, occurs at
annulus velocities of 4 to 5 ft/s9,12, which are at the upper limit
of the flow rates encountered in CTD. It appears though that
the critical flow rate for bed erosion is seldom sufficient to
maintain the particles in suspension5. These authors finally
conclude that
hole cleaning in bed transport is inefficient
a suspension condition that relies only on bed erosion will
under predict the flow rate necessary for efficient
suspension, and hence cuttings removal.
If the flow in the annulus is laminar, the cuttings will
inevitably form a bed, given sufficient length5. Since in CTD

SPE 81746

there is no pipe rotation, the role of turbulence must be

examined for maintaining particles in suspension5,9,12. The
question then posed is whether the flow rate necessary to
suspend the particles in turbulent flow is practical in view of
the limitations on maximum flow rate and pressure drop
imposed for CTD. Leising and Walton5 predicted that fluids of
low to moderate viscosity (5 15 cp at nominal shear rate of
170s-1) were optimal for hole cleaning. A full model, however,
for predicting these critical velocities in terms of the flow
parameters for full suspension of particles is not given.
Minimum Suspension Velocity for Horizontal Flow
Most of the work done and reported in the literature for the
minimum suspension velocity of solids in conduits is for the
pipe geometry, since this is the geometry used for solids
transport using liquids (mainly water)20,29. In addition, much of
the reported work is for high solids loading, up to 40 50%,
with different particle sizes, from fines to coarse particles (up
to 0.275in ). The differences of the studied situations with the
ones we encounter during drilling applications, especially for
drilling with coil tubing are:
the geometry is annulus (most often eccentric)
the inlet concentration of solids is fairly low, rarely
exceeding 4% by volume
the fluid used could be non-Newtonian, shear thinning,
many times modeled as power law fluid
there is a limitation on the maximum flow rate to be
achieved with the upper limit defined by the capacity of
downhole motor and maximum surface pressure
sustained5.

Based on work on solids transport in pipes20,26, the critical

velocity for solids suspension can be derived from
consideration of forces acting on the particles and the
requirement that the forces balance so that particles remain in
suspension. Very relevant is the work of Davies48 where for
full suspension of solids, but not necessarily for the fully
syspended symmetric flow pattern, the balance of forces is
between:
the downward sedimentation force, Fds, and,
the upward eddy fluctuation force, Fue = eddy pressure *
area
These forces are estimated as:
Fds =

d 3p g (1 C s )

Fue = (u ')2 d 2p / 4

(32)
(33)

Hindered settling has been taken into account through the term
(1 C s ) .

When these forces balance, we get

u ' = 0.82 (1 C s )

d p g /

(34)

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 7

The fluctuating velocity (u) for liquid only is related to

energy dissipation from dimensional considerations and
turbulence theory and the fact that we are concerned with
eddies of the size of the particle diameter, hence,

(u ')

= PM d p

(35)

where PM is the power dissipated per unit mass of fluid, given

for a pipe by,
PM =

3
2 f sU M
d

(36)

derived from dimensional considerations. Using Blasius

relationship, for Newtonian fluids,
fs =

0.079

(37)

0.25
N Re,
s

we get finally

( )1 / 3 (d )0.42

u ' = (0.16)1 / 3 ( / )1 / 12 (U M )0.92 d p

(38)

the equation being dimensional and valid for pipe flow and
only Newtonian liquid flowing. Davies48 claims that the
presence of solids dampens u ' and he takes this into account
by letting,
u' p =

u'

(39)

1 + C s

were is a constant. This is in agreement with some investigators26,49-50 but contradictory to others. For example, Julian and
Dukler51, as mentioned in20 studied vertical gas solid flow in
pipes (with results equally applicable to horizontal flows of
solid liquid mixtures). They state that for dilute gas solid
systems the solids make their influence felt by modifying
local turbulence in the gas phase, increasing turbulence
fluctuations, mixing length, eddy viscosity and hence frictional
pressure drop, p . This is also evident from measured curves
of pressure drop which is higher in the presence of solids than
for pure gas or liquid flowing16-17.
Equations (35) and (39), after some algebra, give,
U M = (1.08 )(1 + Cs )

1.09

( / )

0.09

(d )

(1 Cs )

0.46

0.55

2 g

(d )
p

0.54

0.18

(40)

with the equation being dimensional. Utilizing the observed

maximum versus concentration, Davies estimates the value of
= 3.64 . This final equation is then similar to Durands
correlation52, apart from the concentration dependence. The
value of () is set from the equations for hindered settling,
taken as = 4 for 1 < NRe,p < 10 falling to ~ 3 when NRe,p
approaches 100. Davies final predictions were close to
published experimental data but about 1.35 times higher.
If we replace, d, with, dh, and keep everything else the same,
for some typical values, dp = 0.079 in., p = 12.5 ppg, =

8.33 ppg, dh = 2.0 in., Cs = 2%, = 15 cp, = 3, we estimate

UM = 5.1 ft/s!
The above mentioned approach has not been used neither for
the flow in an annulus nor for nonNewtonian fluids, except
in53. The main challenges are then
1. to relate (u) to main flow velocity (UM) for an annulus
and for non-Newtonian fluids. It can be done through the
use of annulus hydraulic diameter and a generalized
Reynold number
2. to examine the dampening or not of the non-Newtonian
liquid velocity fluctuations by the presence of solids and
quantify it
3. to examine the effect of hindered settling for solids
4. to compare predictions with good quality data
5. to verify whether values of UM so predicted conform to
the maximum flow rate and pressure drop imposed for
CTD.
Experimental System
In order to shed more light into efficient cuttings transport for
CTD and aiming at providing good quality data, we
established an experimental system for studying the solids
carrying capacity of liquids in horizontal and inclined annulus,
with no inner pipe rotation.

Before embarking into the setting up of the experimental

system, a dimensional analysis was done to determine the
significant parameters affecting the process. If we visualize the
process for the two layer model (Newtonian liquid solid
suspension and a stationary bed), the height of the bed, h , will
depend on the following parameters:
d p , d o , di , U M , u p , , p , , g

Dimensional analysis suggests equation (41) below,

a

d d 0 b U M (d 0 d i ) d
h

d p d 0 d i

d 0 di

f
k
e
3
2
u p p (d 0 d i ) g

(41)

U
2

M

Hence, the main parameters are, geometrical parameters, d0, di,
dp, dh, physical properties, p , , , velocity ratio (up / UM),

the Reynolds number NRe. The last term in the r.h.s., can be
considered the Galileo number based on hydraulic diameter,

NGa =

d h3 2 g

(42)

with

d h = d 0 di

(43)

Hence, for full simulation of field conditions, we should have

between our system and field conditions not only geometric
similarity, i.e. similar values for d 0 /( d 0 d i ) , d 0 / d p ,
d 0 d i , but also dynamic similarity, i.e. similar values for
velocity ratio (~ volumetric concentration), the Reynolds

V. C. Kelessidis, G. E. Mpandelis

number and properties (Galileo number).

Literature review of SPE papers describing case studies with
CTD, since its first application in 1991 to date4,54-66 gave us
the data of the relevant parameters and are shown in Table 1.
Typical flow rates range between 2 5 bpm, which for typical
annulus sizes results in the velocity of the fluid in the annulus
in the range of 1 to 5 ft/s. It is unfortunate that some
information is missing, as it was not stated explicitly. In
particular, the viscosity of the fluids used, while density was
mentioned and ranged from that of water to a maximum of 9.5
ppg. Based on the above values, we have constructed a flow
loop, shown in Figure 4.
The annulus, made of plexiglass, is 16.4 ft long, with outer
diameter d0 = 2 in. and inner diameter di = 1.57 in. and
currently is concentric. It is supported by a metal structure that
can be tilted from horizontal to vertical. A plastic tank holds
132 gal of liquid and is equipped with a 2hp variable speed
agitator. The flow is achieved with a 10 hp slurry pump
delivering 185 gpm at 64 psig. The flow rate, the density and
temperature are monitored with a Coriolis mass flow meter
(Rheonik, RHM 30). A Validyne pressure transducer monitors
frictional pressure drop in the annulus measuring station at two
axial spacings (15.75 in. and 35.43 in.). Viscosity of liquids is
measured with a continuously variable shear rate coaxial
cylinder viscometer (Grace, type M3500a) with shear rates
continuously varying from 0.01 to 600 RPM. Spherical glass
beads of different diameters (0.04 in., 0.08 in., 0.16 in.) will be
used to simulate drilled cuttings.
A comparison of the pertinent variables between the
experimental system and field operations is given in Table 2.
In Figure 5 we show photos of layer patterns taken from the
preliminary test runs.
Conclusions
A critical review has been presented on the effects of the
various parameters on efficient cuttings transport in horizontal
concentric and eccentric annuli. The modeling and
experimental results show that:
the most significant parameter is the annulus mixture
velocity (flow rate and cross sectional area)
flow should be turbulent in the annulus
maximum flow rates should conform to maximum rates
imposed by downhole motor and maximum pressure
allowable for Coiled Tubing
liquid density is an important parameter
eccentricity plays a significant role with a dramatic
decrease in cuttings transport efficiency
for fully
eccentric annulus
modeling shows that rheology has a small effect in
contradiction to experimental and field results.

From the published results and models to date, the main issues
are the following:
the need for collecting good quality cuttings transport
data, for concentric and eccentric annulus, with conditions

SPE 81746

similar (dynamic similarity) to the ones observed in the

field
the establishment, theoretically, of the observed link (in
the field and experiments) between fluid rheology and
efficient cuttings transport
the establishment of the best rheological models for the
liquids, two versus three parameter models, since the
added complexity of three parameter models is still
questionable
resolving the contradiction for using low viscosity fluids
versus moderate viscosity fluids, where simulation results
show better suspension characteristics for the latter but the
former promote turbulence
to better understand the solids distribution in the
suspension layer. The dispersion coefficient of solids is
very critical and better predictions are needed, especially
for non-Newtonian fluids
should the approach be the determination of minimum
suspension velocity or the modeling of layers for efficient
cuttings transport ?
for the modeling of minimum suspension velocity, the
needs are:
o to find relationships to link the turbulent fluctuating
velocity component to the main flow parameters for
annulus and non-Newtonian fluids
o to get better relationships for hindered settling of
solids in non-Newtonian fluids
o to examine the effect of the presence of solids on the
turbulent fluctuating velocity for non-Newtonian
liquids
o to compare predictions with good quality data
for the layer modeling, the needs are for
o justification of the use of three layers (more complex)
vs the two layer model with good quality data
o better prediction of the solids dispersion coefficient,
D, hence determination of Peclet number
o validation of the diffusion equation for the annulus
o better relationships for the interfacial friction factor,
between heterogeneous layer and moving bed
o better relationships for the wall friction factors for the
heterogeneous layer and for the moving bed, valid for
non-Newtonian fluids.

Appendix A Volumetric Concentration of Solids

During CTD
The solids concentration in the liquid is fairly low during
normal drilling operations, and for CTD rarely exceeds 2 4%
by volume. It can easily be estimated from the following
C =

2
( ROP) * ( / 4 * d hole
) * (1 )

2
Qm + ( ROP) * ( / 4 * d hole
) * (1 )

(A-1)

It should be noted that in the many studies referenced herein,

the bed porosity is rarely taken into account.
For the various field parameters shown in Table 1, taking two
values of d hole (5 in. and 3 in.), we can easily compute the
volumetric concentration of solids, C, for various flow rates

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 9

and we find that the volumetric concentration of cuttings can

vary from about 0.1% to a maximum of 3.5% for the observed
parameters in practice.
Appendix B: Concentration of Solids in the
Heterogeneous Mixture
Many researchers use the solid diffusion equation for
predicting the concentration of solids in the heterogeneous
mixture for two or three layer modeling. This equation is
derived on the assumption that turbulent eddies counterbalance
the settling of solids due to gravity.

With reference to Figure B-1 and using 1-D modeling, with

y the vertical coordinate, a balance between rate of settling
and upward transport by turbulent forces results in4,11,16
up

dC
d 2C
+D 2 =0
dy
dy

(B-1)

The equation is solved using two boundary conditions,

Boundary Condition 1,
no flux of solids at the top wall,

[ u pC ]y = d

dC

+ D
=0
dy y = d

(B-2)

Assuming that D and u p do not vary in the vertical direction,

the solution is,
(B-4)

dyo = d o cos o d o / 2

(B-9)

yi = d i sin i / 2

(B-10)

xi = d i cos i / 2

(B-11)

dyi = d i cos i d i / 2

(B-12)

Ri
Ro

As = 2
xo dyo
xi dyi =
( Ro h)
( Ro h )

/2
/2

1 2
2
2
2
d o (cos o ) d o d i (cos i ) d i
2 ( )

( i )
O

(B-13)
The mean concentration is
Cs =

1
C ( y )dA
As A
s

Cs =

CB 2
d o * Into d i2 * Inti
2 As

Into =

(B-5)

/2

(B-14)

sin o
h
(cos o ) 2 d o

exp Peo

2
d

/2

sin i
h
Int i = exp Peo

2( d o / d i ) d o
i

the above give,

y
h
C ( y)

= exp Peo

CB

d o d o

(B-8)

where,

Using (B-3) and defining the Peclet number, Peo, as,

u p do

xo = d o cos o / 2

and finally,

and using (B-2) we get, A = 0.

Peo =

(B-7)

(B-3)

with CB normally taking the value of 0.52, assuming cubic

packing of the solids.

upy

C ( y ) = A + B exp

yo = d o sin o / 2

The suspension area, As is given by

Boundary Condition 2,

at y = h, C = C B

Let x, y the horizontal and vertical coordinates. Let also

o the angle above horizontal of any point on the perimeter of
the outer pipe and i on the perimeter of the inner pipe (Fig.
B-1 and Fig. 2). Then,

(B-6)

valid for the values of y > h .

(B-14a)

(cos i ) 2 d i (B-14b)

Equations (B-14, B-14a and B-14b) are equations (5, 5a, 5b)
given in the main article.
For these solutions it has been implicitly assumed that both the
settling velocity, u p , and the solids dispersion coefficient, D ,

Equation (B-6) describes the distribution of the particles in the

suspension layer.

are independent of the vertical distance, hence u p is the

The mean concentration of the solids in the suspension is

derived as follows4:

terminal, not the hindered terminal velocity of solids. Doron et

al.16 and Walton4, although they assumed this, they later use
the hindered settling velocity which is obviously not correct,
as already pointed out11.

10

V. C. Kelessidis, G. E. Mpandelis

Estimation of Pe Number
It has been shown above that the concentration of solids in the
heterogeneous layer depends on Peclet number, Pe , defined
by equation (B-5). For Pe << 1 , the process is diffusion
(dispersion) controlled and we expect a more uniform
distribution of solids in the heterogeneous layer. For Pe >> 1 ,
the process is gravity controlled and we expect to see the
solids accumulating near the bottom of the annulus. Therefore,
for full suspension of particles, the process should be diffusion
controlled, or Pe should be minimum.

To get the minimum value of Pe that could be encountered

during CTD, u p and d o should be minimum, while D should
be maximum. We take d p , min 0.0394in. From the proposed
correlations for the terminal velocity the minimum uP is of the
order of 0.0262ft/s, while for d0 the minimum is ~ 2 in.
(Table 1).
For the dispersion coefficient D, there are two approaches
suggested so far in the literature, covering the case of
Newtonian liquids:
APPROACH of Doron et al.16 :
The authors suggest that
D = 0.052(U * )( R )

(B-15)

where U* is the friction velocity given by

U* = U M

fs / 2

(B-16)

and R is the hydraulic radius, which for a full pipe (no bed) is
equal half the pipe diameter.
Extending this to the annulus, we can write,
D = 0.052(U * )( Dh / 2)

(B-17)

using the annulus hydraulic diameter.

The annulus velocity, UM, varies from 1.3 to 4.9 ft/s (Table 1).
For Newtonian fluids it has been suggested5 that optimum ~
10 20 cp.
0.079
For turbulent flow, f s = 0.25 .
N Re s
For D to be maximum, UM should be maximum and fs
maximum or NRe,s minimum. NRe,s is minimum at the onset of
turbulent flow (~ 2100) and hence, fmax = 0.0116.
Hence,
Dmax = 0.052(U max f max / 2 )( Dh / 2) max
2.9 * 10 3 ft 2 / s

APPROACH of Walton4 :
The author suggests that
/3
D = 0.014 D0 * d p * u p * N 1Re,
s

where D0 is given as

(B-18)

SPE 81746

C
D0 = 1.24

0.12
C
D0 =

0.12

0.5

C < 0.05

if

(B-19)

0.25

if

C > 0.05

(B-20)

Equation (B-19) is the case for CTD, and for C = Cmax = 0.05,
(D0)max ~ 0.8.
For this situation, for D to be maximum, D0 should be
maximum (0.8), (dp)max ~ 0.236in., (up)max ~ 0.65 ft/s and
N Re,s max ~ 10000 (from values from Table 1), hence
Dmax 3 * 10 3 ft 2 / s , which is the same order of
magnitude as the approach of Doron et al16.

Thus, the minimum value of Pe expected for annulus for

typical applications of CTD is of the order of
Pe ~

(0.0026 ft / s )(2.75 / 12 ft )
3 * 10 3 ft 2 / s

~3

It is worthwhile to examine the concentration distribution for

various values of Pe. Using equation (B-6) we calculated the
values of C(y)/CB for various values of Pe. The results are
shown in Figure B-2 for h/do = 0.1.
The solids may be distributed close to the pipe bottom at the
minimum estimated value of Pe (~ 3). The effect is more
pronounced for Pe ~ 5. For Pe ~ 1, there is a good distribution
of solids throughout the channel.
The above analysis holds for Newtonian fluids. It shows that
the value of Pe is very critical for determining the solids
distribution in the suspension and the main parameter for
estimating Pe is the solids dispersion coefficient. Furthermore,
it has been shown that no consensus has been reached on the
proper correlation for the solids dispersion coefficient, even
for Newtonian fluids. And no such correlation has been
established for solids dispersion coefficient for non-Newtonian
fluids, which is the case encountered in CTD.
Nomenclature
AB = cross sectional area occupied by the bed, L2, in.2
AM = total cross sectional area, area of annulus, L2, in.2
As = cross sectional area occupied by the suspension
layer, L2, in.2
C = solid volumetric concentration
CB = mean concentration of solids in bed
CM = mean feed concentration
Cs = mean concentration of solids in the suspension
layer
d = pipe diameter, L, in.
dh = hydraulic diameter of annulus, L, in.
dhole = open hole diameter, L, in.
d = diameter of inner pipe of annulus, L, in.
d0 = diameter of outer pipe of annulus, L, in.
dp = spherical particle diameter, L, in.
2
dp = pressure drop due to friction, M/Lt , psi

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 11

D=
DhB =
Dhs =
f=
fB =
fi =
fs =
FB =
FmB =
Fmax =
FN =
Fds =
Fue =
g=
h=
K=
Kv =
n=

dispersion coefficient of solids, L2/t, ft2/s

hydraulic diameter of the bed, L, in.
hydraulic diameter of the suspension, L, in.
friction factor
friction factor for bed wall
friction factor for bed suspension interface
friction factor for the suspension wall
friction force between the bed and the wall, ML/t2,
lbf
friction force between the moving bed and the wall,
ML/t2, lbf
maximum sliding force, ML/t2, lbf
normal force on bottom wall, ML/t2, lbf
downward sedimentation force, ML/t2, lbf
upward eddy fluctuatiung force, ML/t2, lbf
acceleration of gravity, L/t2, ft/s2
bed height, L, in.
Consistency index, M/Ltn, lbfsn/100ft2
von Karman constant
power law index

NGa = Galileo number =

d h3 2 g

NRe = Reynolds number

NRe,s = Reynolds number of suspension, =
NRe,B = Reynolds number of bed, =

NRe,p = particle Reynolds number =

ROP =
Ri =
Ro =
S=
SB =
Si =
Ss =
Qm =
u =

sU s Dhs
s

n
8 sU s2 n Dhs

2(3n + 1)
K

u p d p

Rate of penetration, L/t, ft/hr

radius of inner pipe, L, in.
radius of outer pipe, L, in.
wetted perimeter, L, in.
wetted perimeter of bed, L, in.
wetted perimeter of interface, L, in.
wetted perimeter of suspension, L, in.
mud flow rate, L/t3, gpm
fluctuating component of transverse velocity, L/t,
ft/s
presence of particles, L/t, ft/s
hindered settling velocity of solids, L/t, ft/s
settling velocity of solids, L/t, ft/s
mean velocity of the bed, L/t, ft/s
mixture velocity, L/t, ft/s
moving bed velocity, L/t, ft/s
mean velocity of suspension, L/t, ft/s
vertical coordinate, L, in.
horizontal coordinate, L, in.

= as in Figure 2
= liquid viscosity, M/Lt, cp
= fluid density, M/L3, lbm/gal
= density of bed, M/L3, lbm/gal
= solid particle density, M/L3, lbm/gal

3
s = density of suspension, M/L , lbm/gal

= wall stress between bed and walls, M/Lt2, psi

i = stress at the interface between suspension and bed,

M/Lt2, psi
= wall stress between suspension and walls, M/Lt2,

si

psi
= intergranular shear stress, M/Lt2, psi

s = coefficient of internal friction between particles

u 'p = fluctuating component of transverse velocity in the

uh =
up =
UB =
UM =
UmB =
Us=
y=
z=

= formation porosity (here taken as 20%)

p = sphericity of particles

BU B DhB
B

NRe,gn = Generalized Reynolds number =

Greek Letters
3
= difference between solid and liquid density, M/L ,
lbm/gal
= dry friction coefficient between particles and walls
= as in Figure 2

Subscripts
B = bed
hB = hydraulic, bed
hs = hydraulic, suspension
i = inside, interface
M = mixture
mB = moving bed
mBsB = moving bed stationary bed
o = outside
p = particle
s = suspension
sB = stationary bed
smB = suspension moving bed
References
1. Pilehvari, A. A., Azar, J. J., Shirazi, S. A.: State of the Art
Cuttings Transport in Horizontal Wellbores, paper SPE 37079,
presented at the 1996 SPE International Conference On
Horizontal Well Technology, Calgary, Canada, Nov. 18 20.
2. Pilehvari, A. A., Azar, J. J.: State of the Art Cuttings Transport
in Horizontal Wellbores, SPE Dril. & Compl. (1999), 14 (3),
196 200.
3. Azar, J. J., Sanchez, R. A.: Important Issues in Cuttings
Transport for Drilling Directional Wells, paper SPE 39020,
presented at the 1997 5th Latin American and Caribbean
Petroleum Engineering Conference and Exhibition, Rio de
Janeiro, Brazil, Aug. 30 Sept. 3.
4. Walton, I. C.: Computer Simulator of Coiled Tubing Wellbore
Cleanouts in Deviated Wells Recommends Optimum Pump Rate
and fluid Viscosity, paper SPE 29491, presented at the 1995
Productions Operations Symposium, Oklahoma City, OK, USA,
Apr. 2 4.
5. Leising L. J., and Walton, I. C.: Cuttings Transport Problems
and Solutions in Coiled Tubing Drilling, paper IADC / SPE

12

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.
17.
18.
19.
20.
21.

V. C. Kelessidis, G. E. Mpandelis

39300, presented at the 1998 IADC / SPE Drilling Conference,

Dallas, Texas, USA, Mar. 3 6.
Li, J. and Walker S.: Sensitivity analysis of Hole Cleaning
Parameters in Directional Wells, paper SPE 54498, presented
at the 1999 SPE / ICoTA Coiled Tubing Roundtable, Houston,
Texas, USA, May 25 26 .
Walker S., and Li, J.: The Effects of Particle Size, Fluid
Rheology, and Pipe Eccentricity on Cuttings Transport, paper
SPE 60755, paper presented at the 2000 SPE / ICoTA Coiled
Tubing Roundtable, Houston, TX, Apr. 5 6.
Li, J. and Walker S.: Coiled Tubing Wiper Trip Hole
Cleaning in Highly Deviated Wellbores, paper SPE 68435,
presented at the 2001 SPE / ICoTA Coiled Tubing Roundtable,
Houston, Texas, USA, Mar. 7 8.
Cho, H., Shah, S. N., Osisanya S. O.: A Three Layer Modeling
for Cuttings Transport with Coiled Tubing Horizontal Drilling,
paper SPE 63269, presented at the 2000 SPE Annual Technical
Conference and Exhibition, Dallas, Texas, USA, Oct. 1 4.
Cho, H., Shah, S. N., Osisanya S. O.: Selection of Optimum
Coiled tubing Parameters Through the Cuttings Bed
Characterization, paper SPE 68436, presented at the 2001 SPE
/ ICoTA Coiled Tubing Roundtable, Houston, Texas, Mar. 7 8.
Kamp A. M. and Rivero S.: Layer Modeling for Cuttings
Transport in Highly Inclined Wellbores, paper SPE 53942,
presented at the 1999 Latin American and Caribbean Petroleum
Engineering Conference and Exhibition, Caracas, Venezuela,
Apr. 21 23.
Nguyen, D. and Rathman, S. S.: A three layer Hydraulic
Program for Effective Cuttings Transport and Hole Cleaning in
Highly Deviated and Horizontal Wells, SPE Drilling &
Completion (1998), 13 (3), 182 189.
Martins, A. L. and Santana, C. C.: Evaluation of Cuttings
Transport in Horizontal and Near Horizontal Wells A
Dimensionless Approach, paper SPE 23643, presented at the
1992 Latin American Petroleum Engineering Conference, II
LACPEC, Caracas, Venezuela, March 8 11.
Santana, M. and Martins, A. L.: Advances in the Modeling of
the Stratified Flow of Drilled Cuttings in High Angle and
Horizontal Wells, paper SPE 39890, presented at the 1996
International Petroleum Conference and Exhibition, Mexico,
March 3 5.
Masuda, Y., Doan, Q., Oguztoreli, M., Naganawa, S.,
Yonezawa, T., Kobayashi, A., Kamp, A.: Critical Cuttings
Transport Velocity in Inclined Annulus: Experimental Studies
and Numerical Simulation, paper SPE / Petroleum Society of
CIM 65502, presented at the 2000 SPE / Petroleum Society of
CIM on Horizontal Well Technology, Calgary, Alberta, Nov. 6
8.
Doron, P., Garnica, D. and Barnea, D.: Slurry Flow in
Horizontal Pipes: Experimental and Modeling, Int. J. Mult.
Flow (1987), 13 (4), 535 547.
Doron, P. and Barnea, D.: A Three Layer Model for Solid
Liquid flow in Horizontal Pipes, Int. J. Mult. Flow (1993), 19
(6), 1029 1043.
Doron, P. and Barnea, D.: Pressure Drop and Limit Deposit
Velocity for Solid Liquid Flow in Pipes, Chem. Eng. Science
(1995), 50 (10), 1595 1604.
Doron, P. and Barnea, D.: Flow Pattern Maps for Solid
Liquid Flow in Pipes, Int. J. of Multiphase Flow (1996), 22 (2),
273 283.
Govier, G. W. and Aziz, K.: The Flow of Complex Mixtures in
Pipes, R.E. Krieger Pub. Co, Florida (1972).
Martins, A. L., and Santana, C. C.: Modeling and Simulation of
Annular Axial Flow of Solids and NonNewtonian Mixtures,
paper presented at the 1990 III Encontro Nacional de Ciencias
Termicas, Itapema, SC, Brazil.

SPE 81746

22. Gavignet A. A. and Sobey, I. J.: Model Aids cuttings Transport

Prediction, J. of Petrol. Techn. (1989), Sept., 916 921.
23. Martins, A. L., Santana, M. L. and Gaspari, E. F.: Evaluating
the Transport of Solids Generated by Shale Instabilities in ERW
Drilling, SPE Drill. and Completion (1999), 14 (4), 254 - 259.
24. Silva, M. G. P., and Martins, A. L.: Evaluation of the
Rheological Behavior, Equivalent Diameter and Friction Loss
Equations of Drilling Fluids in Annular Flow Conditions, paper
presented at the 1988 1st World Conference on Experimental
Heat Transfer, Fluid Mechanics and Thermodynamics,
Dubrovnik, Yugoslavia.
25. Televantos, Y., Shook, C. A., Carleton A. and Streat, M.: Flow
of Slurries of Coarse Particles at High Solids Concentrations,
Can. J. Chem. Engr. (1979), 57, 255 262.
26. Cheremisinoff, N., P. and Gupta, R.: Handbook of Fluids in
Motion, Ann Arbor Science, Michigan (1983).
27. Martins, A. L., Sa, C. H. M. and Lourenco, A. M. F.:
Experimental Determination of Interfacial Friction Factor in
Horizontal Drilling with a Bed of Cuttings, paper SPE 36075
presented at the 1996 Fourth SPE Latin American and
Caribbean Petroleum Engineering Conf., Port of Spain, Trinidad
& Tobago, Apr. 23 26.
28. Bagnold, R., A.: Experiments of a gravity free dispersion of
large solid spheres in a Newtonian fluid under shear, Proc. R.
Soc (1954), A225, 49 63.
29. Bagnold, R., A.: The flow of cohensionless grains in fluids,
Phil. Trans. R. Soc. (1957), A249, 235 297.
30. Santana, C. C., Massarani, G., and Sa, C.H.M.: Dynamics of
Solid Particles in NonNewtonian Fluids: The Wall and
concentration Effects, presented at the 1991 22nd Annual
Meeting of the Fine Particle Society, USA, July 31 Aug. 2.
31. Doan, Q. T., Oguztoreli, M., Masuda, Y., Yonezawa, T.,
Kobayashi, A., and Kamp, A.: Modeling of Transient cuttings
Transport in Underbalanced Drilling, paper IADC / SPE
62742, presented at the 2000 IADC / SPE Asia Pacific Drilling
Technology, Kuala Lumpur, Malaysia, Sept. 11 13.
32. Chhabra, R. P., and Peri, S. S.: Simple Method for the
Estimation of Free Fall Velocity of Spherical Particles in
Power Law Liquids, Powder Techn. (1991), 67, 287 290.
33. Machac, I., Ulbrichova, I., Elson, T. P. and Cheesman, D. J.:
Fall of Spherical Particles through NonNewtonian
Suspensions, Chem. Engr. Science (1995, Shorter
Communications), 50 (20), 3323 3327.
34. Miura, H., Takahashi, T., Ichikawa, J., and Kawase, Y.: Bed
Expansion in LiquidSolid Two-Phase Fluidized Beds with
Newtonian and Non-Newtonian Fluids over the Wide Range of
Reynolds Numbers, Powder Techn. (2001), 117, 239 246.
35. Ceylan, K., Herdem, S., and Abbasov, T.: A Theoretical Model
for Estimation of Drag Force in the Flow of NonNewtonian
Fluids Around Spherical Solid Particles, Powder Techn.
(1999), 103, 286 291.
36. Felice, R. Di.: The Sedimentation Velocity of Dilute
Suspensions of Nearly Monosized Spheres, Int. J. Mult. Flow
(1999), 25, 559 574.
37. Tomren, P. H., Iyoho, A. W., and Azar, J. J.: Experimental
Study of Cuttings Transport in directional Wells, SPE Drill.
Engr. (1986), Febr., 43 56.
38. Taylor, G.: The Dispersion of Matter in Turbulent Flow
through a Pipe, Proc. R. Soc. (1954), A223, 446 468.
39. Mols, B. and Oliemans, V. A.: A turbulent Diffusion Model for
Particle Dispersion and Deposition in Horizontal Tube Flow,
Int. J. Mult. Flow (1998), 24 (1), 55 75.
40. Larsen, T. I., Pilehvari, A. A. and Azar, J.J.: Development of a
New Cuttings Transport Model for High Angle Wellbores
Including Horizontal Wells, paper SPE 25872, presented at the

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 13

41.

42.

43.

44.

45.

46.

47.
48.
49.
50.
51.
52.
53.
54.

55.

56.
57.

58.

1993 SPE Rocky Mountain Regional / Low Permeability

Reservoir Symposium, Denver, Colorado, Apr. 12 14.
Larsen, T. I., Pilehvari, A. A. and Azar, J.J.: Development of a
New Cuttings Transport Model for High Angle Wellbores
Including Horizontal Wells, SPE Drill. and Completion (1997),
Vol. 12 (2), 129 135.
Jalukar, L. S., Azar, J. J., Pihlevari, A. A. and Shirazi, S. A.:
Extensive Experimental Investigation of the Hole size Effect
on Cuttings Transport in directional Well Drilling, paper
presented at the 1996 ASME Fluids Engineering Division
Annual Summer Meeting, San Diego, California, July 7 12.
Wilson, K. C. and Tse, J. K. P.: Deposition Limit for Coarse
Particles Transport in Inclined Pipes, Proc. 9th Intl. Conf. on
Hydraulic Transport of Solids in Pipes (1987), BHRA Fluid
Engr, Cranfield, UK, 149.
Ford, J. T., Peden, J. M., Oyeneyin, M. B., Gao E., Zarrough,
R.: Experimental Investigation of Drilled Cuttings Transport in
Inclined Boreholes, paper SPE 20421, presented at the 1990
65th SPE Annual Technical Conference, New Orleans, LA,
USA, Sept. 23 26.
Peden, J. M., Ford, J. T. and Oyeneyin, M. B.: Comprehensive
Experimental Investigation of Drilled Cuttings Transport in
Inclined Wells Including the Effects of Rotation and
Eccentricity, paper SPE 20925 presented at the 1990 Europec,
The Hague, Netherlands, Oct. 22 24.
Chien, S. F.: Settling Velocity of Irregularly shaped Particles,
paper SPE 26121, presented at the 1994 69th Annual Technical
Conference and Exhibition, New Orleans, Louisiana, Sept. 25
28.
Thomas, D. G.: Transport characteristics of Suspensions:
Relation of Hindered Settling Flow Characteristics to
Rheological Parameters, AIChE J. (1963), 9, 310 316.
Davies, J. T.: Calculation of Critical Velocities to Maintain
Solids in Suspension in Horizontal Pipes, Chem. Engr. Science
1987, 42 (7), 1667 1670.
Hsu, F-L., Turian, R., M. and Ma, T-W.: Flow of Noncolloidal
Slurries in Pipelines, AIChE J. (1989), 35 (3), 429 442.
Pechenkin, M., V.: Experimental Studies of Flows with High
Solid Particle Concentration, Proceedings of the 1972 Cong.
A.I.R.H., Tokyo.
Julian, F. M. and Dukler, A. E.: AIChE J. (1965), 11, 853.
Durand R.: Basic Relationships of the Transportation of Solids
in Pipes Experimental and Modeling, Proc. of the 1953 5th
Minneapolis Int. Hydraulics Conv., Minneapolis, MN, 89 103.
Campos, W.: Mechanistic Modeling of Cuttings Transport in
Directional Wells, Ph.D. Thesis (1995), Univ. of Tulsa.
Ramos Jr., A. B., Fahel, R.A,,Chaffin, M. and Pulis, K., H.:
Horizontal Slim-Hole Drilling With Coiled Tubing: An
Operators Experience, paper SPE 23875, presented at the
1992 lADC/SPE Drilling Conference held In New Orleans,
Louisiana, February 10-21.
Leising, L. J. and Newman, K. R.: Coiled Tubing Drilling,
paper SPE 24594, presented at the 1992 67th SPE Annual
Technical Conference and Exhibition, Washington, DC, Oct. 5
7.
Leising L. J. and Rike, E. A.: Coiled Tubing Case Histories,
paper IADC / SPE 27433, presented at the 1994 IADC / SPE
Drilling Conference, Dallas, TX, Feb. 15 18.
Goodrich, G.T. , Smith, B E. and Larson, E.B.: Coiled Tubing
Drilling Practices at Prudhoe Bay, paper SPE 35128, presented
at 1996 lADC / SPE Drilling Conference held in New Orleans
Louisiana, 12-15 March 1996.
Weighill, G., Thereby, H. and Myrholt, L.: Underbalanced
Coiled Tubing Drilling Experience on the Ula Field, paper SPE
35544, presented at the 1996 European Production Operations
Conference and Exhibition, Stavanger, Norway, Apr. 16 17.

59. Elsborg, C., Catter, J. and Cox, R.: High Penetration Rate
Drilling with Coiled Tubing, paper SPE 37074, presented at the
1996 SPE International Conference on Horizontal Well
Technology held in Calgary, Canada, Nov. 18-20.
60. McGregor, B., Cox, R. and Best, J.: Application of Coiled
Tubing Drilling Technology on Deep Under Pressured Gas
Reservoir, paper SPE 38397, presented at the 1997 2nd North
American Coiled Tubing Roundtable, Montgomery, Texas, Apr.
1 3.
61. Svendsen, ., Saasen, A., Vassy, B.,Skogen E., Mackin, F. and
Normann, S. H.: Optimum Fluid Design for Drilling and
Cementing a Well Drilled with Coiled Tubing Technology,
paper SPE 50405, presented at the 1998 SPE International
Conference on Horizontal Well Technology held in Calgary,
Alberta, Canada, Nov. 14.
62. Kara, D.T., Gantt, L.L., Blount, C.G. and Hearn, D.D.:
Dynamically Overbalanced Coiled Tubing Drilling on the
North Slope of Alaska, paper SPE 54496, presented at the 1999
SPE / ICoTA Coiled Tubing Roundtable held in Houston,
Texas, May 2526.
63. Kirk, A. and Sembiring, T.: Application of C.T.D. Offshore,
Indonesia Phase One Pilot Project, paper SPE 54502, presented
at the 1999 SPE/ICoTA Coiled Tubing Roundtable held in
Houston, Texas, May 2526 .
64. Stiles, E. K., DeRoeun M. W., Terry, I. J., Cornell, S. P. and
DuPuy, S. J.: Coiled Tubing Ultrashort-Radius Horizontal
Drilling in a Gas Storage Reservoir: A Case Study, paper SPE
57459, presented at the 1999 SPE Eastern Regional Meeting
held in Charleston, West Virginia, Oct. 2122.
65. Portman, L.: Reducing the Risk, Complexity and Cost of
Coiled Tubing Drilling, paper IADC / SPE 62744, presented at
the 2000 IADC/SPE Asia Pacific Drilling Technology held in
Kuala Lumpur, Malaysia, Sept. 1113.
66. McCarty, T., M. and Stanley, M. J.:Coiled Tubing Drilling:
Continued Performance Improvement in Alaska, paper SPE /
IADC 67824, presented at the 2001 SPE/IADC Drilling
Conference, Amsterdam, 27 Feb. 1 Mar.

TABLE 1: Field Data for CTD Parameters

SOURCE

dhole
(in)

SPE 23875
SPE 24594
" "

di

dh

(in)

(in)

(bpm)

(ft/s)

2.000
4.750

1.750

1.90
3.000

2.00

1.77

4.750

2.000

2.750

3.20

2.95

1.900

3.100

2.38

1.90

5.000

1.900

3.100

4.76

3.81

SPE 35128

3.750

2.375

1.375

SPE 35544

3.500

2.875

0.625

" "

SPE 37074

(ft/hr)

(ppg)

9 to 12

5.000

SPE 29491

ROP

30

8.33
8.33

10 to 70 9.16
7 to 25

2.875

75 to 200

SPE 38397

4.750

2.375

2.375

3 to 50

SPE 50405

3.750

2.375

1.375

40 to 50 12,5

SPE 54496

3.750

2.000

1.750

1.50

2.56

SPE 54502

6.125

2.375

3.750

2.50

1.35

SPE 57459

4.750

2.375

2.375

SPE 62744

2.750

2.000

SPE 67824

2.750

" "
SPE 68436
" "

9.16
20

9.16

20
12 to 21

4.125

12 to 21

5.000

1.900

3.100

1.25

1.00

9.16

5.000

1.900

3.100

6.23

4.99

9.16

14

V. C. Kelessidis, G. E. Mpandelis

SPE 81746

TABLE 2: Comparison of field conditions with experimental set up

Parameter
dhole (in)
di (in)
di / do
(do di) / do
Q (bpm)
U (ft/s)
(ppg)
p (ppg)
dp (in)
C (% by volume)

w

a. SUSPENDED SYMMETRIC

c. MOVING BED

Field Conditions
2.75 3.75 4.75
2 2.375 2.875
0.40 0.53 0.63 0.82
0.37 0.47 0.50 0.61
1.5 - 3.0
1.0 4.9
8.33 9.5
20.8
0.04 to 0.275
0.8 to 4
water & polymers

Experimental Conditions
2.75
1.575
0.57
0.43
0.13 4.4
1.0 14.5
8.33 8.75
20.8
0.04 - 0.08 - 0.16
0.8 to 4
CMC water solutions

60 180

50 - 1800

-1

(= 12U/do-di) (s )

b. SUSPENDED ASYMMETRIC

d. STATIONARY - MOVING BED

Figure 1. Flow patterns for solid liquid flow in horizontal concentric annulus

Figure 2. Schematic for the two layer model

Figure 3. Schematic of the three layer model

SPE 81746 Flow Patterns and Minimum Suspension Velocity for Efficient Cuttings Transport in Horizontal and Deviated Wells in Coiled Tubing Drilling 15

Figure 4. Schematic of annulus flow loop: 1 annulus section, 2 measuring section, 3 tank, 4 agitator, 5 pump,
6 Coriolis flow meter, 7 pressure transducer, 8 P/C for data acquisition

a. Two layers

b. Three layers
Figure. 5: Photos of the layer patterns

0,8

y / do

0,6

0,4

0,2

0
0

0,2

0,4
Pe = 3

Figure B-1: Schematic for deriving diffusion equation

C / CB
Pe = 1

0,6

0,8

Pe = 5

Figure B-2: Solid distribution for h/do=0.1