Академический Документы
Профессиональный Документы
Культура Документы
+
j
(
,
\
,
(
1
1
8 66
1 39
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
The slug holdup for viscosity greater than 500 cp is determined
using a correlation obtained in experiments carried out a 2-in.-test-
loop facility at PDVSA Intevep. Lubrication oil and air were the
testing fluids, and lubrication-oil viscosities from 500 to 1,300 cp
were used. Data of slug-flow characteristics are collected through a
combination of high-speed video camera in the viewing section and
fast signal for pressure drops and wall pressure fluctuations. The
holdup is measured with a set of quick-closing valves. The correla-
tion is based on the Shoham (2000) model, and it is given by
H e
LLS
R
eL
( )
1 0046
0 0022
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
The holdup in the Taylor bubble (the film zone; see Fig. 2) is
determined using the equation obtained by Shoham (2000) as
H
v v H
v
LTB
TB L LLS
TB
( )
. . . . . . . . . . . . . . . . . . . . . . . . . . (7)
The gas-void fraction is the fraction of a volume element in
the two-phase-flow field occupied by the gas phase in the slug
zone. It is expressed using the following equation proposed by
Beggs (1991):
S LLS
H 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (8)
The film length of the slug unit (Fig. 2) is predicted by use of
a correlation developed from experiments conducted with a 2-in.-
test-loop facility at PDVSA Intevep using lubrication oil (480 cp)
and air as testing fluids. Slug-flow-characteristics data are acquired
in the same manner as for Eq. 6. This correlation is based on the
Taitel and Barnea (1990) model and is given by
L
F
SL
SL SG
+
j
(
,
\
,
(
0 0365
0 8606
.
Re
Re Re
.
. . . . . . . . . . . . . . . . . . . (9)
The slug-length (Fig. 2) correlation is obtained by inserting L
F
into the correlation developed by Shoham (2000) to predict the film
length. This correlation considers only hydrodynamic slug flow.
L
v H
S
SL
SL SG
L LLS
+
j
(
,
\
,
(
0 0365
0 8606
.
Re
Re Re
.
vv
SL
1
. . . . . . . . . . . . . . . . . . . (10)
The next correlations require the estimations of velocities, such
as mixture velocity and liquid, and gas superficial velocities. Also,
prediction of the translational velocity (Taylor-bubble velocity) and
the drift velocity is required. Translational velocity is composed of
a superposition of the bubble velocity in stagnant liquid (i.e., the
drift velocity and the maximum velocity in the slug body), and it
is given by the Nicklin (1962) correlation as
v cv v
TB M D
+ , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (11)
where c is derived according to the flow type. If the flow is laminar,
c 2. If it is turbulent c 1 2 . . If flow lies between laminar and
turbulent, the Xiao et al. (1990) correlation is used:
c
L
CL
CL
L
+
j
(
,
\
,
(
+
+
j
(
,
\
,
(
2 0
1
1 20
1
2 2
.
Re
Re
.
Re
Re
. . . . . . . . . . . . . . . . . . . . . . (12)
The drift velocity is the velocity of a phase relative to a surface
moving at the mixture velocity and is calculated depending on the
pipe inclination, as follows.
For pipes slightly inclined, the drift velocity is estimated by the
Bendiksen (1984) correlation as
v v v
D D D
( ) ( ) +( ) ( )
horizontal vertical
cos sin . . . . . . . . . . . . . . (13)
The following correlation is used for horizontal pipes:
v gD
D
( )
horizontal
0 54 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (14)
And for vertical pipes,
v gD
D
( )
vertical
0 35 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (15)
The slug frequency is the rate of recurrence of the slug through
the pipelines and is estimated using the correlation proposed by
Colmenares et al. (2001) as
f
v
L
S
TB
U
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (16)
where slug unit length is given by
L L L
U S F
+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (17)
Other parameters of interest for slug-flow characterization are
the liquid and gas instantaneous flow at the inlet of the catcher.
They are calculated using the Miyoshi et al. (1988) model.
For the liquid,
Q v A H
p LLS insL Mins
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (18)
And for the gas,
Q v A H
p LLS insG Mins
( ) 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . (19)
Prediction of Liquid Accumulation. The liquid accumulation
in the slug catcher is estimated by applying a liquid mass bal-
ance between the inlet and outlet of the equipment (Sarica et al.
1990), as
Liquid-input
mass rate
Liquid-discharge
,
,
]
]
]
mmass rate
Liquid-accumulation
mass rate
,
,
]
]
]
,,
,
]
]
]
.
. . . . . . . . . . . . . . . . . . . . . . . (20)
Slug zone Film zone
V
LLS
V
GLS
V
GTB
V
LTB
H
LTB
L
U
L
F
L
S
V
TB
Fig. 2Schematic for a slug unit (Colmenares et al. 2001).
196 December 2010 SPE Projects, Facilities & Construction
The liquid-input mass rate is determined through the Miyoshi
et al. (1988) model, to calculate the liquid instantaneous flow,
whereas the liquid-discharge mass rate is related to the outlet liquid
flow that depends on the flow-control-valve size.
On the basis of the liquid mass balance presented by Sarica et
al. (1990), the accumulated liquid volume is given as,
V t Q
L
v
v H A Q
sp
S
TB
M LLS p accum acum dis
,
]
]
max
. . . . . . . . . . . . (21)
Finger-Type-Slug-Catcher Sizing. The most important param-
eter in the slug-catcher design is the diameter of the slug-catcher
ngers, which is calculated to obtain stratied ow. In this sense,
models to predict the transition from slug ow to stratied ow are
necessary, such as the IKH instability criterion, the VKH instability
criterion, and the Taitel and Dukler (1976) model.
To predict the transition, Sarica et al. (1990) used the IKH
criterion presented by Taitel and Dukler (1976). However, in this
work it is proposed to use the VKH criterion presented by Barnea
and Taitel (1993) to determine the transition from slug flow to
stratified flow because it better predicts transition for a larger range
of viscosities (100 to 5,000 cp). The criterion is expressed as
v K R R g
A
A
h
Gtran
V L G G L
L G
L G
P
L
+ ( )
j
(
,
\
,
(
LL
,
,
,
,
,
]
]
]
]
]
]
1 2 /
. . . . . . . . . (22)
In this expression, K
V
is a correction factor given as
K
C C
g
A
A
h
V
V IV
L G P
L
L
( )
1
2
cos
d
d
. . . . . . . . . . . . . . . . . . . . (23)
The VKH criterion provides minimum diameter from which
stratification is obtained. When the actual gas velocity is less than
the transition gas velocity, stratified flow is expected. Thus, the
catcher diameter should be bigger than the minimum diameter to
receive the incoming liquid. The catcher diameter is determined by
increasing minimum diameter until stratified flow into the equip-
ment is ensured. Also, one must consider the incoming-liquid flow,
available space for installation, and the costs.
For a given gas superficial velocity, there is a transition liquid
holdup and an operation liquid holdup. The first is given by the
maximum liquid superficial velocity for stratified flow and is cal-
culate using the VKH criterion. The second is given by the average
operation flow rates of liquid and gas at the slug catcher. The differ-
ence between these two holdups will provide the available volume
to handle the accumulation of liquid in the slug catcher; thus, the
catcher length for the designed diameter is given as
L
V
A H H
L L
finger
accum
finger trans oper
]
]
. . . . . . . . . . . . . . . . . . . . . . (24)
Results and Discussion
The proposed methodology is used to design a finger-type slug
catcher for a heavy-oil field in the Orinoco belt. The designed
catcher was compared economically with the design of a con-
ventional horizontal separator that is commonly used in the field.
The field data for the design are given in Table 4. The calculated
procedure to design the catcher using the methodology presented
in the preceding is explained in following.
Procedure To Design a Finger-Type Slug Catcher. To obtain the
main dimensions of the nger-type slug catcher, a computational
procedure is elaborated that is derived from the equations presented
previously.
1. As input data are required operational conditions, properties
of the fluid and the geometry of the inlet pipeline are required,
such as temperature, pressure, API gravity, gas- and liquid-flow
rate, and pipeline diameter and roughness (Table 4).
2. The gas-liquid-flow pattern into the inlet pipeline of the
slug catcher is determined using the model proposed by Barnea
and Taitel (1993). In this sense, a gas/liquid flow-pattern map is
generated for the operational conditions given in Table 4. In the
map, the only areas represented are intermittent flow, annular flow,
and stratified flow. According to Fig. 3, the flow pattern in the
inlet line (pipe diameter is 10 in.) of the slug catcher is located
in the intermittent-flow area as a yellow point called operation
point (pipeline).
3. As detailed in the mapping (Fig. 3), the flow pattern is slug
flow. Therefore, a slug-catcher design is necessary. Without slug
flow, of course, the slug catcher is not necessary. The procedure to
determine the main dimensions of the slug catcher (diameter and
length of fingers) is the following.
a. Slug-flow characteristics are calculated according to correla-
tions and models discussed in this paper (Eqs. 5 through 17).
b. The diameter and quantity of fingers are assumed such that
diameter must be large enough that the quantity of fingers must
be greater than one. In this scenario, four fingers are considered
and an initial diameter of 10 in.
c. Gas- and liquid-flow rates through each finger are determined
considering an even flow distribution among the fingers.
d. The finger diameter is determined using the VKH crite-
rion through an iterative process in which the finger diameter is
increased until obtaining a stratified flow. Thus, minimum diameter
is obtained when the transition curve between stratified flow and
intermittent flow is reached. The point that represents this condi-
tion is shown superimposed on the transition curve in the Fig. 4
(green point), and it is called transition point. The minimum
diameter is increased to the next pipeline commercial diameter to
guarantee stratified flow. In this scenario, the operation point (blue
point) of the finger-type slug catcher is shown in Fig. 4, which
corresponds to a finger-design diameter calculated at 20 in.
e. The catcher length is determined by use of the accumulated
liquid volume and calculated diameter. In this scenario, the length
is 26.2 ft.
f. Finally, the catcher weight is determined considering all pipe
sections, such as inlet header, separation area, gas-outlet header,
and liquid-outlet header. The finger-type slug catcher weight is
7,887 lbm.
Economic Comparison. This subsection presents an economic
comparison between the designed catcher and a conventional
horizontal separator that is commonly used in eld applications
and is sized under the same conditions as those of the catcher. It is
necessary to clarify the cost estimation. It should be used only as
a reference point because it is made for a conceptual design.
Cost estimation is based on fabrication cost because opera-
tion costs are considered comparable because they operate under
the same working principle (gravitational sedimentation), use the
TABLE 4FIELD DATA
6 1 I P A
Temperature (F) 7295
Pressure (psig) 105.80
Q
L
5 5 . 3 4 3 , 4 1 ) D P B (
Q
G
6 5 5 , 8 ) D F S M M (
5 6 . 2 4 ) % ( W & S B
G
5 5 . 0
0 1 ) . n i ( p D
December 2010 SPE Projects, Facilities & Construction 197
same process-control systems, and will require regular cleaning
because of solids accumulation. Therefore, the fabrication cost of
the catcher and conventional separator is estimated relating to the
vessel weight and steel price, as Powers (1990) proposed:
Cost Weight kg Steel price USD
estimated
kg ( )
( )
. . . . . . . . . (25)
The horizontal conventional separators diameter is 72 in., and its
length is 20 ft. The separator weight is approximately 10,214 lbm.
Steel price is considered approximately 30 USD. Using this
referenced price and Eq. 25, the cost of each separator is estimated
(Table 5).
According to Table 5, the fabrication cost of the finger-type
slug catcher is 23% less than that of the horizontal conventional
separator.
Conclusions
On the basis of the methodology of criteria and solutions weight-
ing proposed by Vilchez (2008), the finger-type slug catcher was
selected and designed as the separation technology for heavy-oil
fields.
The slug-flow characteristics must be known to carry out a
proper design for a finger-type slug catcher. Thus, various cor-
relations and models are selected in a rigorous manner to predict
slug-flow characteristics for heavy oil.
VKH HL/D=0.5 Operation point
Intermittent
10
10
VsG [ft/s]
V
s
L
[
f
t
/
s
]
100
1
1
0.1
0.1
0.01
Stratified
Annular
Fig. 3Flow-pattern map for the inlet conditions of the catcher.
Fig. 4Flow-pattern map for the designed slug catcher.
Intermittent
10
10
VsG [ft/s]
V
s
L
[
f
t
/
s
]
100
1
1
0.1
0.1
0.01
Stratified
Annular
VKH
Operation point
Transition point
HL/D=0.5
198 December 2010 SPE Projects, Facilities & Construction
This work proposed an improvement on the Sarica et al.
(1990) methodology on the basis of the use of the VKH criterion
to predict the stratified/no-stratified transition in a more rigorous
way to determine the dimensions of a finger-type slug catcher to
handle viscous liquids. This improvement allows performing a
better design of the slug catcher to guarantee the segregation and
separation of the phases while slug mitigation is achieved for a
heavy-oil field.
Economic comparison demonstrates that the slug catcher costs
approximately 23% less than the conventional separator. Therefore,
it could be used as a separator in heavy-oil fields. However, further
studies conducted in laboratory-scale tests and in fluid-dynamics sim-
ulations have to be conducted before field applications are feasible.
Nomenclature
API API gravity
A cross-sectional area, m
2
C wave velocity
f
s
slug frequency, slugs/s
g gravity acceleration, m/s
2
H
LLS
liquid holdup in the slug zone
H
LTB
liquid holdup in the lm zone (Taylor Bubble zone)
K
v
coefcient of stability
L Length, m
Q ow rate, m
3
/s
Re Reynolds number
t time, s
velocity, m/s
V volume, m
3
Subscripts
accum accumulation
D drift
dis discharge
f liquid lm (Taylor bubble zone)
G gas
ins instantaneous
IV inviscid
M mixture gas-liquid
max maximum
oper operational
p pipe
S slug zone
sG supercial gas
sL supercial liquid
TB translational or Taylor-bubble zone
trans transition
U slug unit
V viscous
Greek Letters
specic gravity
inclination angle
density [kg/m
3
]
Acknowledgment
The authors want to express our gratitude to Joe Bradford and
Maite Bradford (Gazprom Latin America) and Alexis Gammiero
(PDVSA Intevep) for their help and collaboration in the structur-
ing of this work.
References
Barnea, D. 1990. On the effect of viscosity on stability of stratified
gasliquid flowapplication to flow pattern transition at various pipe
inclinations. Chemical Engineering Science 46 (8): 21232131. doi:
10.1016/0009-2509(91)80170-4.
Barnea, D. and Taitel, Y. 1993. Kelvin-Helmholtz stability criteria for strati-
fied flow: viscous versus non-viscous (inviscid) approaches. Int. J. Mul-
tiphase Flow 19 (4): 639649. doi: 10.1016/0301-9322(93)90092-9.
Barnea, D. and Taitel, Y. 1994. Interfacial and structural stability of sepa-
rated flow. Int. J. Multiphase Flow 20 (1): 387414. doi: 10.1016/0301-
9322(94)90078-7.
Beggs, H.D. 1991. Production Optimization Using NODAL Analysis. Tulsa,
Oklahoma: OGCIPetroskills.
Bendiksen, K.H. 1984. An experimental investigation of the motion of long
bubbles in inclined tubes. Int. J. Multiphase Flow 10 (4): 467483. doi:
10.1016/0301-9322(84)90057-0.
Colmenares, J., Ortega, P., Padrino, J., and Trallero, J.L. 2001. Slug Flow
Model for the Prediction of Pressure Drop for High Viscosity Oils in
a Horizontal Pipeline. Paper SPE 71111 SPE International Thermal
Operations and Heavy Oil Symposium, Porlamar, Margarita Island,
Venezuela, 1214 March. doi: 10.2118/71111-MS.
Gregory, G.A., Nicholson, M.K., and Aziz, K. 1978. Correlation of the liq-
uid volume fraction in the slug for horizontal gas-liquid slug flow. Int. J.
Multiphase Flow 4 (1): 3339. doi: 10.1016/0301-9322(78)90023-X.
Lin, P.Y. and Hanratty, T.J. 1986. Prediction of the initiation of slugs with
linear stability theory. Int. J. Multiphase Flow 12 (1): 7998. doi:
10.1016/0301-9322(86)90005-4.
Manzanilla, C. 2007. Identificacin de tecnologas de separacin gas-
lquido para crudos pesados. Informe de pasanta, Universidad de
Carabobo, Valencia, Venezuela.
Manzanilla, C. 2008. Diseo conceptual de un equipo de separacin gas-
lquido para crudos de alta viscosidad. Trabajo de grado, Universidad
de Carabobo, Valencia, Venezuela.
Miyoshi, M., Doty, D.R., and Schmidt, Z. 1988. Slug-Catcher Design for
Dynamic Slugging in an Offshore Production Facility. SPE Prod Eng
3 (4): 563573. SPE-14124-PA. doi: 10.2118/14124-PA.
Nicklin, D.J. 1962. Two-phase bubble flow. Chemical Engineering Science
17 (9): 693702. doi: 10.1016/0009-2509(62)85027-1.
Powers, M.L. 1990. Analysis of Gravity Separation in Freewater Knock-
outs. SPE Prod Eng 5 (1): 5258; Trans., AIME, 289. SPE-18205-PA.
doi: 10.2118/18205-PA.
Sarica, C., Shoham, O., and Brill, J.P. 1990. A New Approach for Finger
Storage Slug Catcher Design. Paper OTC 6414 presented at the Offshore
Technology Conference, Houston, 710 May. doi: 10.4043/6414-MS.
Shoham, O. 2000. Two-Phase Flow Modeling. Thesis, Department of
Petroleum Engineering. University of Tulsa, Tulsa, Oklahoma.
Taitel, Y. and Barnea, D. 1990. Two phase slug flow. In Advances in Heat
Transfer, ed. J.P. Hartnett and T.F. Irvine Jr., Vol. 20, 83132. San
Diego, California: Academic Press.
Taitel, Y. and Dukler, A.E. 1976. A model for predicting flow regime transi-
tions in horizontal and near horizontal gas-liquid flow. AIChE Journal
22 (1): 4755. doi: 10.1002/aic.690220105.
Vergara, M.A. and Foucart, N. 2007. Selection Slug Catcher Type.
Paper SPE 107293 presented at the Latin American and Caribbean
Petroleum Engineering Conference, Buenos Aires, 1518 April. doi:
10.2118/107293-MS.
Vilchez, N. 2008. Estrategias creativas para el diseo mecnico. Disert-
acin, Escuela de Ingeniera Mecnica, Universidad de Carabobo,
Valencia, Venezuela.
TABLE 5FABRICATION-COST ESTIMATION
Separator Type Weight (lbm) Cost (USD)
0 0 . 8 1 8 , 5 3 2 0 0 . 7 8 8 , 7 r e h c t a c g u l s e p y t - r e g n i F
Conventional separator 0 0 . 2 8 3 , 5 0 3 0 0 . 4 1 2 , 0 1
December 2010 SPE Projects, Facilities & Construction 199
Xiao, J.J., Shonham, O., and Brill, J.P. 1990. A Comprehensive Mechanistic
Model for Two-Phase Flow in Pipelines. Paper SPE 20631 presented at
the SPE Annual Technical Conference and Exhibition, New Orleans,
2326 September. doi: 10.2118/20631-MS.
Jos Mrquez is currently a research and development support
engineer at PDVSA Intevep, Los Teques, Venezuela. He joined
PDVSA Intevep in 2002 and has been working on the develop-
ment of phase-separation technologies and multiphase-flow
modeling. He works for the flow assurance and phase sepa-
ration research and development project of PDVSA Intevep,
focusing on multiphase technologies for heavy-oil applica-
tions. Also, he works in the engineering of surfaces facilities of a
pilot project of in-situ combustion. He holds an ME degree from
Universidad de Los Andes (ULA), Venezuela, and an MSc degree
from Universidad Central de Venezuela. Carlos Manzanilla is a
mechanical maintenance supervisor for Nestle of Venezuela. He
worked for PDVSA Intevep in the flow assurance and phase sep-
aration research and development project from 2007 to 2008.
Jorge Trujillo is currently a research and development associate
engineer at PDVSA Intevep, Los Teques, Venezuela. He joined
PDVSA Intevep in 2002 and since then has been working on the
development of high capacity/high efficiency phase-separa-
tion technologies and multiphase-flow modeling. He leads the
flow assurance and phase separation research and develop-
ment project of PDVSA Intevep, focusing on multiphase tech-
nologies for heavy-oil applications. He holds an ME degree from
Universidad Nacional Experimental Politecnica, Venezuela, and
an MSc degree from La Universidad del Zulia, Venezuela.