FLOW REGIME

TRANSmONS

IN IDGH-PRESSURE

INCLINED PIPELINES
1

R. Wilkens!, M. Gopal1, and W. P. Jepson

lSHELL E&P TECHNOLOGY COMPANY

WESTHOLLOW TECHNOLOGY CENTER
3333 IDGHW AY 6 SOUTH
HOUSTON, TX 77082-3101
lNSF IfUCRC, CORROSION IN MULTIPHASE SYSTEMS CENTER
DEPARTMENT OF CHEMICAL ENGINEERING
OIDO UNIVERSITY
. ATHENS, omo 45701

Abstract
An 18-m long, 9.72-cm i.d., inclinable 316 stainless steel pipeline has been commissioned for the
study of multiphase flow and its subsequent effects on corrosion. The effect of inclination and
pressure on flow regime transitions and slug flow characteristics is studied. The fluids used are
carbon dioxide for the gas phase and ASTM substitute seawater with light oil in the liquid phase.
The superficial gas and liquid velocities varied from 1.0-11 mfs and 0.1-1.5 mfs respectively.
Additionally, the pressure varied from 0.27 to 1.48 MPa and the inclination varied from 0 to ::1:5.
The results show that inclination has a dramatic effect on flow regime transitions. Even at an
inclination of +2, the stratified flow has completely disappeared within the range of velocities
studied. Froude numbers are higher at the same velocities in upward inclined flow. Increasing
pressure causes transitions to annular flow at lower gas velocities. Maps are presented for flow
regime transitions at different pressures and inclinations.
Introduction
Previous work at the Corrosion Center (Lee, 1993) has explained multiphase flow regimes and the
mechanisms governing their transitions. A great deal of work has been carried out for two phase
flows in small diameter pipes. This is not scalable to larger pipes. Jepson and Taylor (1993) have
shown that the pipe diameter should be at least 10 em to mimic the mechanisms observed in large
diameter pipelines.
Mandhane et af. (1974) created a two-phase flow map based on the superficial gas and superficial
liquid velocities. Fluid properties, diameter, and inclination specify the flow map which applies.
The Mandhane plots have become a standard format for publishing flow regime data in multiphase
flow.
The first realistic two-phase mechanistic flow regime transition model was produced by Taitel and
Dukler (1976). For the transition from stratified to intermittent or annular flow using the Taitel and
Dukler model, the simultaneous solution of two relations is required, namely, the combined

. .'

momentum balance equation for gas and liquid, and the instability criteria. This model has been
verified (Barnea et al., 1980, etc.) for small-diameter, low pressure, two phase systems at horizontal
to near-horizontal pipe flow. The model has been shown to not work well if the diameter is large
(Jepson and Taylor, 1993) or with the presence of a third phase (Lee, 1993).
Lin (1985) reported large and small diameter flow regime maps for horizontal air-water flow.
Similarly, Jepson and Taylor (1993) and Wallis and Dobson (1973) reported large diameter flow
regime maps, but they were also for horizontal air-water systems. Lee (1993) reported the flow
regime transitions for a large diameter pipe with horizontal three phase flow. This data was for
carbon dioxide gas, water, and a light-oil which is commercially available. Limited flow map data
exists for inclined pipelines. Gould et al. (1974) introduced +45 and +90 flow pattern maps.
Govier and Aziz (1972) presented a commonly used method of establishing flow patterns for
inclined flow. Barnea et al. (1985) proposed a model predicting transitions in inclined pipelines.
Stanislav et al. (1986) reported inclined flow pattern data. Kokal and Stanislav (1986)
characterized, extensively, the upflow and downflow patterns. The models and data compared well,
however all of these studies involved two-phase flow. Additionally, flow in large-diameter pipes
and at high-pressure have not been reported in inclined pipelines. Further, little research has been
done on three phase flow regimes and their transitions. This work will provide the data necessary
in large diameter three phase flow to include the effects of inclination and pressure.
Experimental Setup
An 18-m long, 9.72-cm i.d., high-pressure (13 MPa), high temperature (90C), inclinable 316
stainless steel flow loop has been commissioned for the study of multiphase flow and its subsequent
effects upon corrosion. Figure 1 is a schematic of the system. A predetermined oil and water
mixture is stored within a 1.4 m3 mixing tank. The liquid is moved through the system by a
centrifugal pump powered by a 3 - 15 kW variable speed Baldor motor and its flow rate maintained
by the gate valves labeled A and B. The flow rate is determined with a TMfR 510 frequency
analyzer which was calibrated to a GH Flow Automation (model 6531) in-line turbine flow meter.
A 2-MPa feed line supplies carbon dioxide gas from a 20,000 kg receiver. After passing through
a pressure regulator, the gas flow rate is set by adjusting ball valve C. A Hedland variable area flow
meter is used to determine the gas flow rate. The gas temperature and pressure are monitored
between the flow meter and the pressure regulator. The gas then passes through a check valve, to
avoid possible liquid backflow, and into the liquid flow. The multiphase mixtures then enters the
test loop through a compression flange, allowing the inclination to be set at any angle. Upon
entering the inclined portion of the test loop, the multiphase mixture travels 18 meters before
reaching the test section.
Figure 2 illustrates the test section with the instrument port locations. Port A is a fluid sampling
port used primarily when preparing for corrosion experiments. System temperature is measured
through port B with a type-K thermocouple connected to an OMEGA DP3200- TC electronic
analyzer with display. Any of the ports labeled C can be coupled and used to measure differential
pressure. In these experiments, the differential pressures are measured between the two sets oftaps
placed lOand 132-cmapart. The measurements are made with Oto35 kPaOMEGAPX-750heavy

-'

duty differential pressure transducers. The entire pressure-signal based system has been patented in
a non-visual technique to determine flow regime transition and is not described here. Port D is used
to monitor the test section pressure. This pressure is measured with a 0 to 2.8 MPa Noshok pressure
gauge. The ports marked E can be used to insert corrosion probes if necessary. Additional data can
be taken using two upflow and two downflow acoustic sensors provided by BP Research.
Upon leaving the test section, the multiphase flow passes through a separator to prevent siphoning
due to the declined angle of flow return and to destroy the flow pattern. The mixture passes back
through another compression flange and then re-enters the mixing tank. The gas passes through a
de-entrainment plate through a back-pressure regulating control valve, through a separator, and is
vented to the atmosphere. The liquid from the separator is collected to be re-injected into the
system.

Test Matrix
The matrix studied is listed in Table 1. ASTM Dl14l-52 substitute seawater with an oil of density
800 kg/m3 and viscosity 2 cP were used in the liquid phase with carbon dioxide in the gas phase.
Table 1: Experimental test matrix for flow regime and flow property determination.

property
water cut
pressure
inclination
temperature

range
40,80, 100%
0.27,0.45,0.79
horizontal,

:t:

MPa
2,

:t:

20 DC

diameter

0.0972m

superficial gas velocity

0- 13 mls

superficial liquid velocity

0.1, 0.5, 1.0, 1.5 mls

Results and Discussion

The flow regimes were determined from the criteria established by Wilkens and Jepson (1996). The
flow regimes identified were plug flow, stratified flow, slug flow, pseudo-slug flow, and annular
flow. Plug flow, slug flow, and pseudo-slug flow will often be collectively termed slug flow. Plug
flow is actually of little interest and is not known to occur in downflow. Slug flow was found to
dominate the flow regime map as the inclination was increased to as little as +2. This is expected
as it has been found by many researchers (Kokal and Stanislav, 1989, etc.). Figure 3 is a flow
regime map for 100% saltwater, horizontal, 0.45 MPa flow. At superficial liquid velocities of up
to 0.3 mis, stratified flow is observed to occur while slug flow was observed to occur at a superficial
liquid velocity of 0.4 mls. Pseudo-slug and annular flow occurred at the higher gas flow rates while

. .,

plug flow occurred at the lower gas flow rates. Figure 4 represents the flow regime map for 100%
saltwater, +5 inclined, 0.27 MPa flow. No stratified flow was observed to occur. In its place at
equal flow rates is slug flow. At a superficial liquid velocity as low as 0.1 mfs, slug flow is still
observed to occur, allowing slug flow to dominate the flow regime map.
In downward flow, stratified flow dominates the flow regime map. The transition from stratified
to slug flow also becomes much more dependent upon the superficial gas velocity. Figure 3 showed
that the transition from stratified to slug flow on the axes given was relatively horizontal (i.e.,
occurring at a similar superficial liquid velocity for all superficial gas velocities studied). Figure
5 shows that if the pipe inclination is set to _2, the transition becomes much more dependent upon
the gas flow rate. At a superficial gas velocity of about 1 mfs, only stratified flow is observed at
superficial liquid velocities as high as 1.5 mfs. At a superficial gas velocity of about 3 mIs, slug
flow occurs at a superficial liquid velocity as low as 1 mfs while stratified flow occurs at a
superficial liquid velocity of 0.5 mfs. At a superficial gas velocity of around 9 mfs, slug flow is
observed to occur at a superficial liquid velocity as low as 0.5 mfs while stratified flow occurs at a
superficial liquid velocity of 0.1 m/s. This trend is observed at other water cuts and at other
pressures.Figure 6 shows that with 80% water cut, the transition from stratified to slug flow does
not changes greatly compared to 100% water cut in Figure 5.
These stratified-slug transition results are expected and have been seen by other researchers (Kokal
and Stanislav, 1989, etc.). In downflow, the liquid film is thinner and faster. At lower superficial
gas velocities, more liquid is required to bridge across the pipe. At higher gas velocities, the film
thickness is not much different from that in horizontal flow at high gas velocities, and the transition
occurs near where it is expected to occur in horizontal flow. Further downward inclination causes
the transition from stratified to slug flow to occur at higher liquid flow rates.
The transition to annular flow was found to occur in roughly the same location for all conditions
tested at a superficial gas velocity around 10 mfs. The transition occurred at lower gas flow rate
with low liquid flow rates and at a higher gas flow rate for the higher liquid flow rates. This was
also observed by Kokal and Stanislav (1989). Inclination was found to have little effect on the
transition in the range of conditions tested here. It appears that the gas flow rate required to reach
annular flow is slightly lower in upflow and slightly higher in downflow. But nothing is observed
which exceeds the uncertainties associated with the superficial gas velocity. Kokal and Stanislav
also observed a slight decrease in gas required to reach annular flow with an increase in inclination,
but it was on the order of their uncertainty. They concluded that this transition was relatively
insensitive to inclination (_9 to +9). Water cut was also found to have little observable effect on
the transition to annular flow for the conditions tested.
Pressure was found to have a marked effect on the transition. As the pressure was increased, the
transition to annular flow was observed to occur at lower superficial gas velocities. This effect has
been observed in the field (Green, 1997) and is reasonable. Since annular flow is largely a density
driven effect, it follows that the ratio of densities of the process fluids should affect this transition.
In oil-water flows, when annular flow conditions occur, the less-dense and more viscous fluid (oil)
flows in the core. In gas-liquid annular flow, the less-dense and less viscous fluid (gas) flows in the
core. Since liquid-liquid annular flow occurs at a less-dense fluid superficial velocity of around 1

mfs for an oil with a specific gravity of near unity and around 5 mfs for an oil with a specific gravity

of around 0.8 (Brauner and Maron, 1992), and since for gas-liquid annular flow occurs at a lessdense fluid superficial velocity of around 10 mfs, the closer the densities are, the lower the velocity
requirement. Brauner and Maron demonstrated that as the oil specific gravity approached unity, the
3
effect increased rapidly. As listed earlier, the gas density increases from 5.02 to 14.9 kg/m as the
pressure is increased from 0.27 to 0.79 MPa. Although this is a slight change with respect to the
liquid density, there is a large effect on the ratio of the two.
Figures 4 and 7 represent the same flow conditions at pressures of 0.27 and 0.79 MPa, respectively.
At 0.27 MPa and a superficial liquid velocity of 0.1 mfs, slug flow was observed to occur at a
superficial gas velocity of about 8 mfs. At 0.79 MPa and a superficial liquid velocity of 0.1 mfs,
annular flow was found to occur at a superficial gas velocity as low as 7 mfs. This effect can also
be seen at other water cuts and at other inclinations.
Conclusions
Inclination is found to have a dramatic effect on flow regime transitions. Stratified flow was
eliminated in upflow while slug flow was found to dominate. In downflow stratified flow was
dominant while slug flow was reduced. In downflow, water cut was found to have little measurable
effect on the transition from stratified to slug flow. Water cut was found to have little effect on the
transition from slug to annular flow. Increasing pressure caused the stratified to slug transition to
occur at slightly higher liquid flow rates. The transition from slug to annular flow was found to not
be largely dependent on the inclination. Increasing pressure caused the annular transition to occur
at lower gas flow rates.
References
Bamea, D., Shoham, 0., Taitel, Y., and Dukler, A> E., "Flow Pattern Transitions for Gas-Liquid
Flow in Horizontal and Inclined Pipes: Comparison of Experimental Data with Theory," Int. J
Multiphase Flow, 6, 217-225, 1980.
Gould, T. L., Tek., M., Katz., D. L., "Two Phase Flow Through Vertical, Inclined or Curved Pipes,"
J Petrol. Tech., 26, 915-926, 1974.
Govier, G. W., and Aziz, K., "The Flow of Complex Mixtures in Pipes", Van Nostrand Reinhold.
New York, 1972.
Jepson, W. P., and Taylor, R. E., "Slug Flow and its Transitions in Large-Diameter

Horizontal

Pipes", Int. J Multiphase Flow, 19,411-420, 1993.

Kokal, S. L., and Stanislav, 1. F., "An Experimental Study of Two-Phase Flow in Slightly Inclined
Pipes-I. Flow Patterns, "Chemical Engineering Science, 44,665-679, 1989.
Lee, AJ-Hsin, "A Study of Flow Regime Transitions for Oil- Water-Gas Mixtures in Large Diameter
Horizontal Pipelines", M.S. Thesis, Ohio University, Athens, Ohio 1993.

~.

Lin, P. Y., "Flow Regime Transitions in Horizontal Gas-Liquid Flow", Ph.D. Thesis, University of
Illinois, Urbana-Champaign,

1985.

Mandhane, 1. M., Gregory, G. A., and Aziz, K, "A Flow Pattern Map for Gas-Liquid Flow in
Horizontal Pipes: Predictive Models, " Int. J Multiphase Flow, 1, 537-553, 1974.
Taitel, Y, and Dukler, A. E., "A Model for Predicting Flow Regime Transitions in Horizontal and
Near Horizontal Gas-Liquid Flow", AlChEJ, 22, 47-55, 1976.
Wallis, G. B., and Dobson, J. E., "The Onset of Slugging in Horizontal Stratified Air-Water Flow",
Int. J Multiphase Flow, 1, 173-193, 1973.

test section

separator

LEGEND:

pressure

gauge

temperature

gauge

flow gauge

0
<D
<D

check valve

I<J

gate valve

I><J

ball valve

rupture disk
compression

flange

"

[]X[]

Figure 1: High-pressure, inclinable flow loop orientation.

mixing

centrifugal
carbon dioxide feed line

pump

,,

flow
~

oEE

one

1 D-cm
l32-cm

meter

differential
differential

pressure
pressure

taps
taps

LEGEND:

Figure 2: Diagram of the test section.

void fraction port

thermocouple port

differential pressure tap

system pressure/shear stress port

C
D

corrosion probe insertion port

o
o

+
+

plug

+ slug

pseudo-slug

annular

stratified

0.05
0.5

10

1
superficial gas velocity [=] m/s

Figure 3: Flow regime map for 100% saltwater at 0.45 MPa and horizontal.

20

i.)

'0
qJ

'I

.,

0
1
CIl

IC

+ slug

++

plug

I <>

pseudo-slug

8
.--.
II

'--'

p
.-<

+ +

(.)

0
Q)

:>

.-<
.-<6c.-<a
"d

(.)

'
Q)

gCIl

0.1

0.05

0.5

10

1
superficial gas velocity [=] m/s.

Figure 4: Flow regime map for 100% saltwater at 0.27 MPa and +5 inclination.

20

. .,
~~~ , ...
:.

1
ell

S
,.......,
II
.....

0
....
t)

0
....
~
:>

....
"Cl

6....
....
....
co
....
t)

'~

g.
ell

0.1

0.05
0.5

10

1
superficial gas velocitY [=] m/s

Figure 5: Flow regime map for 100% saltwater at 0.45 MPa and _2 inclination.

slug

stratified

20

~I

tI)

hoI

"

~):

,~

+ +

l-

"-

"

S
,......,
.....

+ +

l-

ell

II
.....

C,)

0
....
Q)

>
....
g.
....
....
ca
.....
"'d

C,)

'
g.
Q)

ell

0.1

superficial gas velocity [=] m/s

Figure 6: Flow regime map for 80% saltwater/20% light oil at 0.27 MPa and _2 inclination.

stratified

10

slug

0.05
0.5

1+

20

..

L,'l.,

JA

"

_) III
~..

II.A ,

t .

.
0

+
+

ell

'8
II
'--'

~
.....
u

:>

.....
.....6t.....a
u
'"d

0::l

ell

0.1

0.05

t
0.5

<>

I
I

IC

plug

I+

slug

I0

pseudo-slug

II
10

annular

20

I
\

superficial gas velocity [=] mls

Figure 7: Flow regime map for 100% saltwater at 0.79 MPa and +5 inclination.
i