Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
∂u v ∂v v 2 w 2 sin α ∂ ⎛ p⎞
u + − − = − ⎜⎜ ⎟⎟
∂r r ∂α r R + r sin α ∂r ⎝ ρ ⎠
Introduction ⎛1 ∂ cos α ⎞⎛ ∂v v 1 ∂u ⎞
−υ⎜ + ⎟⎜ + − ⎟
Coiled tubing (CT) has been used in well drilling, completion, ⎝ r ∂α R + r sin α ⎠⎝ ∂r r r ∂α ⎠ ........................................................(1)
stimulation, wellbore cleanout, and other operations in the petrole-
um industry(1). Accurate prediction of frictional pressure losses
when pumping fluids through coiled tubing has remained a chal-
∂v v ∂v uv w 2 cos α 1 ∂ ⎛ p⎞
lenge in hydraulics design, mainly due to the lack of adequate u + + − =− ⎜ ⎟
∂r r ∂α r R + r sin α r ∂α ⎝ ρ ⎠
friction loss correlations and proper understanding of the complex
flow phenomena of fluids (especially non-Newtonian fluids) in ⎛∂ sin α ⎞⎛ ∂v v 1 ∂u ⎞
+υ⎜ + ⎟⎜ + − ⎟
coiled tubing. Because of the effect of centrifugal forces, sec- ⎝ ∂r R + r sin α ⎠⎝ ∂r r r ∂α ⎠ ..............................................................(2)
ondary flow occurs when a fluid flows through a coiled tubing. It
is a known fact that flow in coiled tubing encounters more pres-
sure losses than in straight tubing. It is believed that secondary
flow causes the excessive friction losses. Since the classic work of
Dean(2, 3), the flow of Newtonian fluids in coiled pipes has been
extensively studied; in contrast, the flow of non-Newtonian fluid
in coiled pipes has remained relatively unstudied.
The objective of this paper is to review both theoretical and
experimental studies on the flow of Newtonian and non-
Newtonian fluids in coiled pipes. The mathematical formulation
and the general characteristics of the secondary flow are first
introduced in order to prepare for discussion of various theoretical
studies. The available friction factor correlations will be compared
and evaluated for their accuracy and applicability.
A recent increase of coiled tubing applications has been driving
the research activities of coiled tubing hydraulics using full-scale
experimental facilities. This paper will also present recent devel-
opments in those experimental investigations.
It is hoped that this comprehensive review and the experimen-
tal results will provide valuable information for those interested in
FIGURE 1: Toroidal coordinate system.
the application and research of coiled tubing hydraulics.
PEER REVIEWED PAPER (“REVIEW AND PUBLICATION PROCESS” CAN BE FOUND ON OUR WEB SITE)
⎡⎛ ∂ 1 ⎞⎛ ∂w w sin α ⎞ 1 ∂ ⎛ 1 ∂w w cos α ⎞⎤
+ υ⎢⎜ + ⎟⎜ + ⎟+ ⎜ + ⎟⎥
⎣⎝ ∂r r ⎠⎝ ∂r R + r sin α ⎠ r ∂α ⎝ r ∂α R + r sin α ⎠⎦
.......................(3)
∂u u u sin α 1 ∂v v cos α
+ + + + =0
∂r r R + r sin α r ∂α R + r sin α ..............................................(4)
1 ∂p
− = G (constant)
R ∂θ ...........................................................................(5)
It can be shown that K and NDe are related by K = 2(NDe)2. Theoretical Methods
Another definition of Dean number(6) is: The complexity of the flow geometries and the equations of
fluid flow in coiled pipes attracted the attention of theoreticians as
⎛ 2 a 3 ⎞ Ga 2 well as engineers. Various theoretical methods have been devel-
D= ⎜ 2 ⎟
⎝ν R⎠ µ oped, which can be roughly grouped as analytical solutions for
.......................................................................................(9)
small Dean number, numerical methods, and boundary layer
methods for large Dean number.
Here, D and K are related by D = 4 K .
The Dean number provides a fundamental parameter in devel-
Analytical Solutions
oping flow resistance correlations for flow in curved pipes. It has
been found that at low Dean number, the law of resistance can be For loosely coiled tubes (a/R is small) and small Dean number,
correlated with NDe only. For high Dean number, both NDe and the governing equations can be simplified and reduced to equa-
curvature ratio (a/R) will be required. tions about a stream-function (ψ) and the axial velocity (w). Since
FIGURE 4a: Secondary flow streamlines and axial velocity FIGURE 4b: Secondary flow streamlines and axial velocity
contours (D = 96)(6). contours (D = 481)(6).
FIGURE 5a: Contours of axial velocity (D = 2,000)(10). FIGURE 5b: Contours of axial velocity (D = 5,000)(10).
⎢ ⎝ R ⎠ ⎥ De ⎝ R ⎠ ⎝R⎠
fCL ⎣ ⎦
= 1+
14 Liu and Masliyah fSL 1 + 49 N De 0 < a/R < 1, NDe < 5,000
when NDe < 100. Interestingly, all four correlations—Ito (for NDe In recent years, the rapid increase of coiled tubing applications
> 100), Collins and Dennis(10), Liu and Masliyah(50), and Mishra in the oil and gas industry has driven the research activities of
and Gupta(44)—are in close agreement over a large range of Dean coiled tubing hydraulics using full-scale facilities. Azouz et al.(54)
number. measured frictional pressure losses of linear guar gum, hydroxy-
A few turbulent correlations are listed in Table 2. There is close propyl guar (HPG), and borate-crosslinked guar gum in 38.1 mm
agreement between them. (11/2 in.) coiled tubing reels. Willingham and Shah(55) reported an
empirical friction factor correlation based on a series of flow tests
with polymer solutions through 25.4, 38.1, and 60.3 mm (1, 11/2,
Non-Newtonian Fluid and 23/8 in.) coiled tubing reels. Their correlation constants are
The complex rheological behaviour of non-Newtonian fluids functions of apparent viscosity and tubing ID.
adds another dimension of complexity to the flow phenomena in McCann and Islas(56) generalized the Srinivasan et al.(43) corre-
coiled pipes. The flow of non-Newtonian fluids in coiled pipes lation for turbulent flow to non-Newtonian fluids and compared
has remained much less studied than Newtonian fluids. the generalized correlation with the full-scale tests of six fluids
Jones(11) considered the theoretical problem of the flow of a prepared using bentonite and lime.
non-Newtonian, visco-inelastic Reiner-Rivlin fluid in a coiled
tube. A number of investigators(12, 13, 51-53) have reported studies Shah and Zhou(45) discussed the characteristics of drag reduc-
on the effect of elasticity on fluid flow in coiled pipes. Jones and tion of polymer solutions in coiled tubing and affecting parame-
Davies’s(52) experiments showed that minute amounts of polymer ters such as curvature, tubing ID, and polymer concentration.
could significantly delay departure of flow rate from Poiseuille More recently, Shah et al.(57) experimentally investigated the flow
flow, a phenomenon called “drag reduction in the laminar region.” behaviour of hydraulic fracturing slurries in coiled tubing. An
Tsang and James(53) noted that polymer additives caused a reduc- empirical correlation of slurry friction factor as affected by sand
tion of the secondary motion and tried to explain the drag reduc- concentration was developed. The possible mechanisms of tubing
tion by estimating the cross-sectional stresses, based on Dean’s erosion in typical fracturing operations have been addressed.
solution and several molecular models.
The present review shows that, compared with its counterpart
Among the various studies of non-Newtonian flow in coiled
of Newtonian fluids, the flow of non-Newtonian fluids has
pipes, those by Mashelkar and Devarajan(37-39) deserve special
remained very much unstudied, either theoretically or experimen-
attention. Following Ito’s (34) boundary layer approximation
tally. Yet, most fluids used with coiled tubing operations in the oil
approach for Newtonian fluids, Mashelkar and Devarajan(37-39)
and gas industry are non-Newtonian, for example, polymer-based
theoretically analyzed and numerically solved the flow equations
solutions, and surfactant-based solutions for drilling, completion,
for a power-law fluid for the conditions of both laminar and turbu-
or wellbore cleanout. Slurries and foam fluids are also used with
lent flows. An empirical correlation for laminar flow was devel-
coiled tubing. The available correlations are not appropriate for
oped based on the numerical solutions and another correlation for
predicting friction pressure losses in these applications. Further
visco-elastic fluids in terms of the Weissenberg number based on
systematic research is required.
data of flow tests.
Mishra and Gupta(44) have also measured friction factors of As with the case of Newtonian fluids, we have summarized the
non-Newtonian fluids in coiled pipes and empirical correlations available friction factor correlations of non-Newtonian fluids in
have been obtained. coiled tubing in Table 3.
⎡
( )
4.0 ⎤ 16
fCL = fSL ⎢1 + 0.033 log N De2 fSL =
⎣ ⎥⎦ N Re 2
2 Mishra and Gupta
12
⎛a⎞ 0.079
fCT = fST + 0.0075⎜ ⎟ fST =
3
⎝ R⎠ 0.25
N Re d
0.1
1.06a ⎛ a ⎞
fCT = 0.8 b ⎜ R ⎟
()
log10 n + 3.93
g ⎝ ⎠
N Re a=
4 McCann and Islas 50
b=
1.75 − log10 n ()
7
) ( )
B d, µ
(
fC = A d , µ @511 +
@ 511
N Re g
5 Willingham and Shah Expressions for A and B are given
in Reference (59)