Copyright 2000, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 2000 SPE Annual Technical Conference and
Exhibition held in Dallas, Texas, 14 October 2000.
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 posi-
tion 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 com-
mercial purposes without the written consent of the Society of Petroleum Engineers is prohib-
ited. 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
The ability to monitoring bottomhole flowing pressure in
pumping oil wells provides important information regarding
both reservoir and artificial lift performance. Converting sur-
face pressure measurements to bottomhole is currently accom-
plished by locating the fluid level in the annulus using a sonic
device and then applying a correlation to estimate the density
of the gas-cut liquid column above the perforations. This gas-
cut liquid column exhibits Zero Net Liquid Flow (ZNLF) con-
ditions where the casing head gas flows through a nearly stag-
nant liquid. A review of literature reveals that the correlations
currently in use by the industry are based on low pressure and
low viscosity data. Increasingly, operators are encountering
high viscosity fluids and in some instances, casinghead gas
flows into gathering systems which are not the typical low
pressure systems. This paper presents recent experimental
measurements for ZNLF under high viscosity and pressure
conditions. The results indicate that viscosity and pressure
have a significant effect on the calculated bottomhole pres-
sures and that these effects are not accounted for in the cur-
rently applied methods. New methods are presented to calcu-
late multiphase density, i.e. gradient correction factor, under
high viscosity and high pressure conditions.
Introduction
The ability to monitor bottomhole pressures provides many
advantages for reservoir management. Pumping wells com-
pleted without a packer provide a special opportunity for this
low cost and reliable bottomhole pressure surveillance. For
these pumping wells, wellhead pressure data is converted to
bottomhole pressure by use of flow models and an acoustic
(sonic) device to locate the gas-liquid interface. Fig. 1 shows
the schematic of this type of pumping oil well. The well is
completed in a conventional fashion, without a packer. The
pump can be a sucker rod pump, PC pump or ESP. The pro-
duced fluids are pump from the well through the tubing string,
while produced and solution gas travels up the tubing/casing
annulus and is produced as casinghead gas at the surface.
Acoustic devises are used to determine the depth to the inter-
face, either from direct velocity measurement as discussed by
Thomas et al.
1
or through correlation of the casing collar re-
flection with the well pipe tally as discussed by McCoy et al.
2
.
Once the gas-liquid interface has been located, bottomhole
pressure is estimated through use of flow models to calculate
the pressure drop through the gas phase above the interface
and oil-water-gas mixture that exists below the interface. This
paper considers the prediction of pressure drop of this liquid-
gas mixture that existing below the interface. In particular,
what effect do fluid viscosity and pressure have on these pre-
dictions.
The flow of gas through a static liquid column creates a spe-
cial type of multiphase flow, termed Zero Net Liquid Flow
(ZNLF). In this case, liquid is present in the wellbore but does
not flow out the tubing/casing annulus with the gas phase.
The gas phase simply passes through the often-churning col-
umn of liquid. A number of methods have been developed to
predict the liquid holdup of this gas-cut liquid column. The
well-known Gilbert chart correlates a Liquid Correction Factor
(LCF), F
x
, with an F
y
term of Q
G
/AP
0.4
. The LCF corrects the
liquid density to account for the effect of the gas phase. In
modern terms, the LCF is related to liquid holdup, H
L
, by:
( )
L G L L m L
H H LCF + 1 ) (
..(1)
and the F
y
term is simply (v
SG
)
sc
/P
0.4
. Apparently the P
0.4
term
is used to approximately convert standard condition gas flow
rate to in-situ conditions, thereby approximating v
SG
. In
1977, Godbey & Dimon
3
presented a correlation of v
SG
with
the gas void faction, 1-H
L
. The most widely use methods was
proposed by Podio et al.
4
in 1980. They found the Gilbert
chart to be valid only at extremely low pressures (0 psig) and
that the Godbey & Dimon did not provide a good match of
their data. They examined pressures from 0 to 50 psig in their
SPE 63047
Estimation of Bottomhole Pressure in Pumping Oil Wells: Effect of High-Viscosity
Fluids and Casinghead Pressure
H. An, SPE, Louisiana State University; S.L.Scott, SPE, Texas A&M University, and J.P. Langlinais, SPE, Louisiana State
University
2 H. AN, S.L. SCOTT, J.P. LANGLINAIS SPE 63047
study and proposed a new modified S curve for ZNLF in
water-air system as shown below

+
8
1
0 0
n
n
n L
X C C H
, .(2)
where C
0
= 6.59329 E-01, C
1
= -4.22842 E-01, C
2
= -1.56392
E-01, C
3
= 4.62897 E-02, C
4
= 6.11411 E-02, C
5
= 1.54776 E-
02, C
6
= -6.51489 E-03, C
7
= -4.58866 E-03, C
8
= -7.12890 E-
04, and X = log
10
(v
SG
).
Their experimental results showed that the increase in liquid
viscosity had an effect on ZNLF holdup, when viscosity was
increased from 1 to 50 cp. The experimental data fitted the
modified S curve, Eq. 2 by defining a corrected v
SG
:
06 . 0
/ '
SG SG
v v
, .(3)
where
cp 50 1
Hasan et al.
5
presented a model for prediction of the LCF
which allows a variety of fluid propoerties to be considered.
In 1994, Kabir & Hasan
6
present a comprehensive review of
the current methods and discuss the accuracy of the various
available methods.
Current Models of Vertical ZNLF
Most models of ZNLF are slip modesl, based on calculating
holdup from the differing velocities of the gas and liquid
phases. The velocity of the gas phase is based on modeling
the movement of a Taylor Bubble:
T m G
v v C v

+
0
.....(4)
This would be expected to valid in the bubble and slug flow
regimes, but also appears to extend into the churn flow region
as well. Using the definition of slip and the fact that the net
liquid velocity is zero, yields:
T SG
SG
S
SG
L
v v C
v
v
v
H

+

0
0
1 1
, ..(5)
where C
0
is the flow distribution coefficient (the ratio of the
centerline velocity to the average velocity of the flux), and v
T
is the Taylor Bubble rise velocity in a stagnant liquid column.
White & Beardmore
7
investigated the effect of viscosity,
density, and surface tension on the Taylor Bubble rise velocity
in a stagnant liquid column. Dimensionless analysis led to the
identification of several key dimensionless numbers such as
Eo (Eotvos number), M (Molton number), and Fr (Froude
number). Froude number, u
2
/(gD), represents the ratio of in-
ertial to gravitational forces. The Molton number, g
4
/(
3
),
does not include either u or D, but contains only the properties
of the fluid. Eotvos number, gd
2
/, has the effect of surface
tension, not viscosity nor velocity, and is needed to describe
the phenomena of Taylor Bubble rise velocity. For viscosities
less than 1000 cp and normal oil field tubulars and fluids, their
results showed little change in drift velocity. If only viscosity
increases, and surface tension, density, Eo, diameter of pipe,
C
0
, and v
SG
are constant, the Froude number decreases. There-
fore, calculated ZNLF holdup, Eq. 5, at constant v
SG
would
tend to decrease slightly with increasing viscosity.
Aziz et al.
8
stated that Taylor bubble rise velocity in a stag-
nant column is
L
G L
T
gD
C v

) (

, ..(6)
where C was given by Wallis
9
as
( )
[ ]
]
]
]
]
,
,

,

(
(
,
\
,
,
(
j
m
N
N
E
v
e e Fr C
37 . 3
029 . 0 5 . 0
1 1 345 . 0
,
( )
L
g L
E
gD
N

2
, ( )
L
g L L
v
g D
N

3
,
and m is determined from
N
v
m
250 10
250>N
v
>18 69N
v
-0.35
18 25
The ZNLF fluid column can be either full pipe flow or flow
in the annular space between the tubing and the casing. Cae-
tano investigated the effect of a concentric and eccentric an-
nulus. He found that large Taylor bubbles did move slighly
faster in an annular geometry than for full pipe flow. Even
though it resulting in overprediction of the Taylor bubble rise
velocity for all his data, Caetano
10
recommended the approach
of Sadatomi et al.
11
whereby the diameter in the bubble drift
equation is replace by the equi-periphery diameter which is the
sum of the casing and tubing diameters. Clearances for fish-
ing operations limit the range of tubing to casing diameters
that occur in oil field completions. In this range, there would
appear to be no difference between flow in an annulus and
pipe flow for the fully eccentric geometry. For concentric
geometry, Caetanos data lie between the pipeflow values and
those predicted using the Sadatomi et al. diameter correction.
Recently, Arpandi et al.
12
presented a model for ZNLF
holdup within a comapct gas-liquid separator:
]
]
]
]
,
,

,
(
(
,
\
,
,
(
j

(
(
,
\
,
,
(
j

P
d
G
SG
L
L
L
v
v
H 1 1
0
0
, (7)
The first term in this equation uses the equation for ZNLF
holdup as shown in Eq. 5. A constant value for C
0
is assumed
as 1.15 for slug/churn flow. The second term is a correction to
account for that liquid holdup will exist only in that portion of
ESTIMATION OF BOTTOMHOLE PRESSURE IN PUMPING OIL WELLS:
SPE 63047 EFFECT OF HIGH-VISCOSITY FLUIDS AND CASINGHEAD PRESSURE 3
pipe that is slug-churn flow. L
P
stands for the length of pipe,
and L
d
does for the length of pipe in annular flow, or the
length of the droplet region. This portion of the pipe will not
be included in the calculation of liquid holdup, since very little
liquid exists in the upper portion of vertical ZNLF. The equa-
tion for the length of the droplet region was derived from a
droplet ballistic analysis using C
d
= 0.44 suggested by Turner
et al.
13
as shown below
( )
(
(
,
\
,
,
(
j

C L
SG G
d
SG
d
g
v
C
v
g
L

32
3
2
2
1
2
2
, .(8)
Experimental Results
Recently, experimental studies have been conducted to exam-
ine the role of high pressure, viscosity and density on liquid
holdup for ZNLF. Duncan
14,15
constructed a 25 ft tall, 4.897-
inch ID high pressure test apparatus and conducted tests with
high pressure methane and water. To investigate effects of
density and viscosity in vertical ZNLF, An
16,17
constructed a
14-ft, 3-in ID vertical transparent pipe apparatus and con-
ducted tests with air and a wide variety of fluids. Liu &
Scott
18
also looked at the effects of viscosity using a 3-inch,
30-ft tall clear acrylic apparatus. This apparatus was also used
to examine the effect of fluid improvement chemicals, such as
drag reducing agents and foamers, on ZNLF
19
. The details of
these experiments and the results obtained are given in Refs.
14-19. In summary, pressure, viscosity and chemicals were all
found to have dramatic effects on liquid holdup for ZNLF.
The water-air experimental data of An
16
at 25 psig was
compared with Podio et al.
4
and Arpandi et al.
12
. The experi-
mental data show an excellent match with the correlation of
Podio et al.
4
(Fig. 2). However, the experimental data exhibit
a discrepancy with the correlation of Podio et al.
4
as the flow
regime reaches annular flow.
Fig. 3 illustrates the effect of density in ZNLF holdup. As
density increases from 6.8 ppg (LVT 200 Oil) to 11.5 ppg
(Calcium Chloride Solution), the ZNLF holdup is observed to
increase. As viscosity increases from 1cp to 100cp (water
through the Glycerin Solution), the ZNLF holdup is signifi-
cantly increased as shown in Fig. 4. Pressure makes a signifi-
cant difference on ZNLF holdup as shown in Fig. 5. The
modified S curve was compared with experimental data as
shown in Figs. 6-7.
From White and Beardmore
7
, the rise velocity of the Taylor
bubble in a stagnant liquid column is unaffected by viscosity if

2
gD
3
/
2
>3*10
5
and surface tension if Eo >70. All the New-
tonian fluids used by An
16
were found to have essentially
identical v
T
.
Modeling and Correlation of ZNLF
The ZNLF slug flow equation, Eq. 5, is used to analyze the
ZNLF phenomena. From the assumption of the identical v
T
from the vertical ZNLF slug to the churn flow, the flow distri-
bution coefficient, C
0
, changes as a function of v
SG
as shown
in Fig. 8. The transition from slug to churn is indicated by
where C
0
has a value of 1.2. Similarly, the transition from
churn to annular is indicated where C
0
takes a value of 1. The
correlation of C
0
vs. v
SG
was made for effects of liquid viscos-
ity, liquid density and pressure. From experimental data, poly-
nomial lines were put into data points. Then, by using Aziz et
al.
8
correlation for Fr
0.5
and changing C
0
, Eq. 5 was put into
polynomial lines. The correlation of C
0
was found as,
) exp(
0
d cv b av C
SG SG
+ + +
, ...(9)
On the graph (Fig. 8) of C
0
vs. v
SG
from LVT 200 oil to cal-
cium chloride (6.8ppg to 11.5ppg), by assuming viscosity ef-
fect is neglected, the correlation of C
0
vs. density was
founded. First of all, from the four straight-line parts from
6.8ppg to 11.5ppg in vertical ZNLF churn flow regime, which
C
0
is less than 1.2, the average slope, a, was calculated. By
trial and error, it was founded that a logarithmic function de-
scribed the relation between C
0
and liquid density well. There-
fore, intersection, b, was put as logarithmic relation between
C
0
and liquid density. However, in vertical ZNLF slug flow
regime, which C
0
is greater than 1.2, different constants had to
be plugged into the exponent of the correlation because verti-
cal ZNLF slug flow regime had a greater density effect on C
0
than vertical ZNLF churn flow regime. In the same manner,
on the graph of C
0
vs. v
SG
from 20cp glycerin (9.7ppg) to
100cp glycerin (11.5ppg), the correlation of C
0
vs. viscosity
was found. Since the density of 50cp glycerin is 10ppg, the
10ppg calcium chloride was the base line for the correlation.
The same method was applied to the 100cp glycerin (11.5ppg)
comparing with 11.5ppg calcium chloride. There were four
points to construct the correlation of C
0
vs. viscosity, which
were 8.33ppg, 9.7ppg, 10ppg, and 11.5ppg. The 9.7ppg line
was from the correlation. It was founded that a linear function
described the relation between C
0
and viscosity well. Intersec-
tion, b, was put as linear relation between C
0
and viscosity.
Like density effect, exponent parts of the correlation had more
viscosity effect on C
0
than the straight-lines. Different slope,
c, and intersection, d, were put into the correlation. From this
regression analysis method, the corrected C
0
for effects of liq-
uid viscosity, liquid density, and pressure was found.
4
'
3 2 1
'
0
ln A A A v A C
G
L
SG
+ +
(
(
,
\
,
,
(
j
+

]
]
]
,

,
+ +
(
(
,
\
,
,
(
j
+ +
8
'
7 6 5
ln A A A v A EXP
G
L
SG

..(10)
where
osity water visc
viscosity liquid
viscosity liquid ess dimensionl
'

,
A
1
= -0.0246, A
2
= 0.1654, A
3
= 0.001, A
4
= 0.399, A
5
= -
0.6757,A
6
= 1.6407, A
7
= 0.0036, and A
8
= -9.4357
4 H. AN, S.L. SCOTT, J.P. LANGLINAIS SPE 63047
By using ZNLF slug flow equation and Aziz et al.s Correla-
tion
8
for Fr
0.5
, vertical ZNLF holdup can be calculated in dif-
ferent liquid density, liquid viscosity, and pressure conditions.
T SG
SG
L
v v C
v
H

+

'
0
0
1
, ..(11)
Figs. 9-11 show that this proposed correlation matches very
well with experimental data.
New Approach for ZNLF Flow Pattern Transitions
Various kinds of mechanistic models exist to predict multi-
phase flow pattern transitions. The slug-churn transition and
churn-annular transition from this experimental data do not fit
any of the existing models. Therefore, a new approach is
needed to describe the phenomena of vertical ZNLF.
In bubble-slug transition, using equation in An et al.
17
, the
transition of air-water system at 25 psig is,
sec / 2026 . 0
) (
3825 . 0
4 / 1
2
ft
g
v
L
G L
SG

]
]
]
,

, .(12)
However, drift flux mechanistic approach in An et al.
2
yields
sec / 2583 . 0
) (
4829 . 0
4 / 1
2
ft
g
v
L
G L
SG

]
]
]
,

, ...(13)
From the correlation of Podio et al.
4
, ZNLF holdup of water-
air at 25psig as 0.75 exists at 0.58 ft/sec. Also, the proposed
correlation, Eq. 9 yields that ZNLF holdup, 0.75 exists at
0.525 ft/sec.
In slug flow regime, there are two parts, which are the Tay-
lor bubble and liquid slug having small bubbles, which chases
the tail of the Taylor bubble. The Taylor bubble moves up-
ward, taking almost the entire pipe area. In another words,
there is a tiny falling film between the Taylor bubble and the
pipe wall. This falling film has an effect on the chasing liquid
slug. Besides, the friction on the pipe also effects this flow
regime. The non-viscous flow regime of the liquid slug is not
affected by this tiny falling film. In ZNLF conditions, liquid is
moving up by gas flow and down by gravity force. The liquid
in one unit slug can be considered as the liquid holdup. How-
ever, the bottom of the pipe has more liquid slugs in vertical
ZNLF than the upper part of the pipe because of gravity force.
But, in a viewpoint of the liquid holdup in the entire pipe
length, we can consider each unit slug contains the same
amount of liquid.
Taitel et al.
20
modeled the slug-churn transition. They as-
sumed that immediately behind the Taylor bubble, liquid
moves up at the same velocity as the Taylor bubble. But, just
before the next Taylor bubble, liquid in the centerline moves
up at the centerline mixture velocity, C
0
v
M
. However, in
ZNLF conditions, since superficial liquid velocity is zero,
C
0
v
SG
can be accepted as the centerline velocity. By assuming
the exponential variation between two liquid velocities, the
slug-churn transition in water-air system was established. The
v
SG
in which churn flow can be observed in the entire length
of the test apparatus (14ft) is,
(
(
,
\
,
,
(
j
+ 292 . 0 6 . 42
gD
v
D
L
SG E
,
sec / 9 . 2 ft v
SG

, ..(14)
The slug-churn transition approach from Taitel et al.
20
dis-
agreed with the transition from the experimental data. The
experimental ZNLF slug-churn transition was set by using C
0
as 1.2 because C
0
is changed by the superficial gas velocity in
vertical ZNLF condition. This transition has 0.261 as liquid
holdup and 6.5 ft/sec as v
SG
in the water-air system under ver-
tical ZNLF conditions. The experimental transition exists at
the much higher v
SG
in water-air systems comparing with
Taitel et al.
20
model. Viscosity and density make the differ-
ence on this transition zone as well as the value of ZNLF
holdup from results of this experiment. For example, in 100cp-
glycerin solution (11.5ppg), the transition exists at 11.9 ft/sec
and ZNLF holdup is 0.221. Another slug-churn transition was
founded by Jayanti and Hewitt
21
at 0.22 as liquid holdup.
Similarly, the experimental ZNLF churn-annular transition
was set at the point, where C
0
becomes to 1. This transition has
0.064 as liquid holdup at 14.5 ft/sec in water-air system in
vertical ZNLF condition. However, for the 100cp-glycerin
solution (11.5ppg), the transition is at above 15 ft/sec from
experimental data as well as the different liquid holdup.
Why Does More Liquid Exist in High Density and
Viscosity Liquids and Less Liquid Exist in High
Pressure in Vertical ZNLF?
In several viscous fluids, transitions indicate different values
where liquid holdups are less than 0.261 because the amount
of liquid slug is getting smaller at high superficial gas veloci-
ties. Since a viscous liquid is hard to break up and sticks to the
pipe side, higher v
SG
is needed to achieve slug and churn flow
regimes. In a vertical viscous Zero Net slug flow, the Taylor
bubble moves up in a more narrow flow area. The smaller
diameter of the Taylor bubble makes more falling film and
moves up much faster. Since more liquid is falling down to the
liquid slug, a thick film develops in the pipe side, and the
centerline velocity (slip velocity) becomes much faster than
the average velocity of the product of flux. It effects C
0
, the
flow distribution coefficient. Therefore, C
0
must increase due
to the thick film falling as viscosity increases. Also, in high-
density fluids, the heavy falling film falls faster to the liquid
slug, which makes C
0
increase. Fig. 12 shows the schematic of
slug flow regimes, which shows the falling liquid film near the
Taylor bubble increases as viscosity increases.
By considering that lengths of a unit slug are the same for
non-viscous and viscous fluids, and diameters of bubbles in
viscous fluids are smaller than in non-viscous fluids and move
faster, higher holdup in high density and viscosity liquid in
vertical ZNLF can be explained. Also, under high pressure
conditions, gas density approaches the liquid density. By
mixing heavier gas with liquid, C
0
is getting flattened, ex-
plaining why holdup decreases under high pressure conditions.
ESTIMATION OF BOTTOMHOLE PRESSURE IN PUMPING OIL WELLS:
SPE 63047 EFFECT OF HIGH-VISCOSITY FLUIDS AND CASINGHEAD PRESSURE 5
Proposed Technique of Estimating Flowing Bottom-
hole Pressure in Pumping Oil Wells
Using the following data, calculate the flowing bottomhole
pressure in the pumping oil well
Casing - 4 in. ID
Tubing - 2 in. OD
Depth to perforation zone - 10000 ft
Producing annulus fluid level - 3000 ft
Producing casing head pressure - 200 psig
Producing casing head temperature - 60 F
Casing head gas rate - 200 Mscfd
Gas specific gravity - 0.8
Annular liquid - oil,
o
= 0.952
Surface tension at standard condition - 60 dynes/cm
Geothermal gradient - 0.010625 F/ft
Total produced GOR - 700 scf/STB
Solution
PVT calculation was from Kartoatmodjo and Schmidt Corre-
lation
22
and Dranchuk and Abou-Kassem Equation
23
of State.
The moody friction equation with equi-periphery diameter was
used in gas column. Using Cullendar and Smith method
24
, the
flowing pressure at the annulus fluid level can be calculated.
For liquid column, pressure traverse calculation and Eq. 9
were used to estimate mixture pressure gradient. Mixture pres-
sure gradient in liquid columns is p/d (psi/ft) = 0.052
M
=
[ ] ) 1 ( 052 . 0
L G L L
H H +
.
The computer program has been made to calculate flowing
bottomhole pressure in pumping oil well by using Visual Ba-
sic in Microsoft Excel. By putting more increments of two
methods, more accurate flowing bottomhole pressure can be
estimated in various densities and viscosites of gas and liquid.
Input data, output data, pressure profile, and data sheet from
this computer program are shown in Figs. 13-16.
Conclusions
1. ZNLF Slug flow equation was used to analyze the vertical
ZNLF phenomena with good results. The correlation of cor-
rected flow distribution coefficient (C
0
) was made to estimate
ZNLF holdup in different liquid densities, viscosites, and
pressures.
2. As density increased from 6.8 ppg (LVT 200 Oil) to 11.5
ppg (Calcium Chloride Solution), the ZNLF holdup increased.
It was founded that a logarithmic function described the rela-
tion between C
0
and liquid density well.
3. As viscosity increases from 1 cp to 100 cp, the ZNLF
holdup is increased significantly. By comparing identical den-
sities of calcium chloride and glycerin solutions, viscosity was
found to have an effect on the vertical ZNLF holdup. It was
founded that a linear function described the relation between
C
0
and liquid viscosity well.
4. As pressure increases from 100 psig to 1000 psig, the
ZNLF holdup is decreased. It was founded that a logarithmic
function described the relation between C
0
and 1/gas density
well.
5. By using this computer program, flowing bottomhole pres-
sure in pumping oil wells can be estimated more accurately in
various liquid densities, viscosites, and pressure conditions.
Nomenclature
a,b,c,d,C = constant
C
0
= flow distribution coefficient
C
d
= drag coefficient
D = pipe diameter
d = depth, ft
g = gravity acceleration
H = holdup
L = length
m = mass
p = pressure, psi
v = velocity, ft/sec
V = volume
X = Constant
= density, pound per gal
= surface tension, dynes/cm
= viscosity, cp
Subscripts and Superscripts
0 = zero net liquid flow conditions
BS = bubble swarm
C = centerline of pipe or gas core in an-
nular flow
d = droplet region
E = entrance
G = gas or gravity
GLS = gas in liquid slug
GTB = gas in Taylor bubble
L = liquid
LCF = liquid correcton factor
LLF = liquid in rising annular film
LLS = liquid in liquid slug
LS = liquid slug
Lo = liquid in falling film
M = mixture of gas and liquid
P = pipe
S = slip
Sc = standard conditions
SG = superficial gas
SL = superficial liquid
SLS = stable liquid slug
SU = slug unit
ULS = unstable liquid slug
T = Taylor bubble in a stagnant liquid
column
Acknowledgement
Hyoung-Jin An wishes to express appreciation to the LSU
Petroleum Engineering Research and Technology Transfer
Laboratory (PERTTL) for their assistance. Also, Mr. An
wishes to express his gratitude to the financial support from
the Louisiana Energy Enhancement Program (LEEP) and the
6 H. AN, S.L. SCOTT, J.P. LANGLINAIS SPE 63047
LSU Petroleum Engineering Department. The authors are
grateful for the efforts of Richard Duncan, Bazil Elzein and
Emily Baie who assisted with this study.
Reference
1. Thomas, L.K., R.W. Hankinson and K.A. Phillips: Determina-
tion of Acoustic Velocities for Natural Gas, SPE J. of Petro-
leum Technology, 889-895 (July 1970).
2. McCoy, J.N., A.L. Podio, K.L. Huddleston and B. Drake:
Acoustic Static Bottomhole Pressures, SPE paper 13810 pre-
sented at the Production Operations Symposium, Oklahoma City
(March 10-12, 1985).
3. Godbey, J.K. and C.A. Dimon: The Automatic Liquid Level
Monitor for Pumping Wells, SPE J. of Petroleum Tech., pp
1019-1024 (August 1977).
4. Podio, A.L., Tarrillion, M. J., and Roberts, E.T.: Laboratory
work improves calculations, Oil & Gas J. (Aug 1980) 137.
5. Hasan, A.R., C.S. Kabir and R. Rahman: Predicting Liquid
Gradient in a Pumping-Well Annulus, SPE Production Engi-
neering, pp 113-120 (Feb. 1988).
6. Kabir, C.S. and A.R. Hasan: Two-Phase Flow Correlations as
Applied to Pumping Well Testing, ASME J. of Energy Re-
sources Tech., Vol. 116, 121-127 (June 1994). October.
7. White, E.T., and Beardmore, R.H.: The Velocity of Rise of
Single Cylindrical Air Bubbles Through Liquids Contained in
Vertical Tubes, Chemical Engineering Science (1961), Vol.17,
351.
8. Aziz, K., Govier, G.W., and Fogarasi, M.: Pressure Drop in
Wells Producing Oil and Gas, J. Cdn. Pet. Tech. (July-
September 1972).
9. Wallis, G.B.: One-Dimensional Two-Phase Flow, McGraw Hill,
New York, NY (1969).
10. Caetano, E.F.: "Upward Vertical Two-Phase Flow Through an
Annulus," Ph.D. Dissertation, The University of Tulsa (1986).
11. Sadatomi, M., Y. Sato and S. Saruwatari: Two-Phase Flow in
Vertical Noncircular Channels, Int. J. Multiphase Flow, Vol. 8,
pp 641-655 (1982).
12. Arpandi, I., Joshi, A.R., Shoham, O., Shirazi, S., and Kouba,
G.E.: Hydrodynamics of Two-Phase Flow in Gas-Liquid Cy-
lindrical Cyclone Separators, paper SPE 30683 Presentated at
the 1998 SPE Annual Technical Conference & Exhibition, Dal-
las, TX, 22-25.
13. Turner, R.G., Hubbard, M.G., and Dukler, A.E.: "Analysis and
Prediction of Minimum Flow Rate for the Continuos Removal
of Liquid from Gas Wells," JPT. (1969), Vol.21, 1475.
14. Duncan, R.W.: Effects of High Pressure on Liquid Holdup for
Vertical Zero Net Liquid Flow, M.S thesis, Louisiana State
University, Baton Rouge, LA (1998).
15. Duncan, R.W., and Scott, S.L.: Vertical Zero-Net Liquid Flow:
Effects of High Pressure on Holdup, paper presented at the
1998 BHRG Multiphase Conference, Banff, Canada, Vol. 31,
43.
16. An, H.: Effects of Density and Viscosity in Vertical Zero Net
Liquid Flow, M.S thesis, Louisiana State University, Baton
Rouge, LA (2000).
17. An, H., Langlinais, J.P., and Scott, S.L.: Effects of Density and
Viscosity in Vertical Zero Net Liquid Flow, paper presented at
the ASME-ETCE Production Technology Symposium, New
Orleans (Feb. 14-17): ASME J. of Energy Resources Tech. (June
2000).
18. Liu, L. and S.L. Scott: A Model for Vertical Zero Net-Liquid
Flow, paper presented at the ASME-ETCE 2000 Production
Technology Symposium, New Orleans (Feb. 14-17, 2000).
19. Lui, L. and S.L. Scott: Effect of Flow Improvement Chemicals
on Vertical Zero Net-Liquid Flow, proceedings of the BHRG
2nd North American Conference on Multiphase Technology,
Banff, Canada (June 21-23, 2000).
20. Taitel, Y., Barnea, D., and Duckler, A.E.: "Modeling Flow Pat-
tern Transitions for Steady State Upward Gas-Liquid Flow in
Vertical Tubes," AIChE J. (1980), Vol.26, 345.
21. Jayanti, S., and Hewitt, G.F.: Prediction of the Slug-To-Churn
Flow Transition in Vertical Two-Phase Flow, Intl. J. Multi-
phase Flow (1992), 6, 847.
22. Kartoatmodjo, T. and Schmidt, Z.: New Correlations for Crude
Oil Physical Properties, paper SPE 23556 available from SPE,
Richardson, TX.
23. Dranchuk, P.M. and Abu-Kassem, J.H.: Calculation of Z-factor
for Natural Gases Using Equations-of-State, J. Cdn. Pet. Tech.
(July-September 1975).
24. Cullender, M.H., and Smith, R.V.: Practical Solution of Gas-
Flow Equations or Wells and Pipelines with Large Temperature
Gradients, Trans., AIME (1956), 207.
SI Metric Conversion Factores
cp x 1.0* E - 03 = Pas
dyne x 1.0* E - 02 = mN
ft x 3.048* E - 01 = m
F (F - 32)/1.8= C
gal x 3.785 412 E - 03 = m
3
in. x 2.54* E + 00 = cm
lbm x 4.535 924 E - 01 = kg
psi x 6.894 757 E + 00 = kPa
ZNLF
Pump
Perforation
Zone
Liquid
Gas
Fig.1. Schematic of Pumping Oil Well
ESTIMATION OF BOTTOMHOLE PRESSURE IN PUMPING OIL WELLS:
SPE 63047 EFFECT OF HIGH-VISCOSITY FLUIDS AND CASINGHEAD PRESSURE 7
0.0
0.7
0 2 4 6 8 10 12 14 16
VSG(ft/sec)
Experimental Data of water
Correlation of Podio et al.
Arpandi's Model
Excellent Match
Fig.2. Comparison of Water-Air System in ZNLF condition
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 2 4 6 8 10 12 14 16 18
VSG (ft/sec)
Z
N
L
F

H
o
l
d
u
p
Oil (6.8ppg)
Water (8.33ppg)
Calcium Chloride (10ppg)
Calcium Chloride (11.5ppg)
C
o
=1
Slug-Churn Transition,C
0
=1.2
Churn-Annular Transition,C
0
=1
Fig.3. Effect of Density on ZNLF Holdup (from An et al.
17
)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
VSG (ft/sec)
Z
N
L
F

H
o
l
d
u
p
20cp Glycerine (9.7ppg)
50cp Glycerine (10ppg)
Calcium Chloride (10ppg)
100cp Glycerine (11.5ppg)
Calcium Chloride (11.5ppg)
Slug-Churn Transition,C
0
=1.2
Fig.4. Effect of Viscosity on ZNLF Holdup (from An et al.
17
)
0.0
0.1
0.2
0.3
0.4
0.5
0 2 4 6 8 10 12 14 16 18
vSG (ft/sec)
100 psig 200 psig
300 psig 400 psig
500 psig 1000 psig
Fig.5. Effect of Pressure on ZNLF Holdup
0.01
0.10
1.00
10.00
100.00
0.0 0.2 0.4 0.6 0.8 1.0
ZNLF Holdup
Oil (6.8ppg)
Water
20 cp Glycerin (9.8ppg)
50 cp Glycerin (10 ppg)
100 cp Glycerin (11.5ppg)
Modifed "S" Curve - Podio et al.
Fig.6. Comparsion with Modified S Curve (Density and Viscosity
Effect)
0.01
0.10
1.00
10.00
100.00
0.0 0.2 0.4 0.6 0.8 1.0
ZNLF Holdup
100 psig
200 psig
300 psig
400 psig
500 psig
1000 psig
Modifed "S" Curve - Podio et al.
Fig.7. Comparsion with Modified S Curve (Pressure Effect)
8 H. AN, S.L. SCOTT, J.P. LANGLINAIS SPE 63047
C
0
= 1.6081V
sg
-0.1942
C
0
= 1.3087e
-0.0208Vsg
C
0
= 1.8839V
sg
-0.2464
C
0
= 1.3897e
-0.0227Vsg
C
0
= 1.9578V
sg
-0.2471
C
0
= 1.4264e
-0.0228Vsg
C
0
= 2.0079V
sg
-0.2517
C
0
= 1.4501e
-0.0233Vsg
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
0 2 4 6 8 10 12 14 16 18 20
Vsg(ft/sec)
C
o
Oil (slug)
Oil (churn)
Water (slug)
Water (churn)
Calcium Chloride 10ppg (slug)
Calcium Chloride 10ppg (churn)
Calcium Chloride 11.5ppg (slug)
Calcium Chloride 11.5ppg (churn)
Fig.8. C0 vs. vSG in ZNLF Holdup (from An et al.
2
)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 2 4 6 8 10 12 14 16
V
SG
(ft/sec)
H
L
0
Oil (6.8 ppg)
Water
Calcium Chloride (10 ppg)
Calcium Chloride (11.5 ppg)
Fig.9. Proposed Correlation of ZNLF Holdup (Density Effect)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 2 4 6 8 10 12 14 16
V
SG
(ft/sec)
H
L
0
20 cp Glycerin (9.7 ppg)
50 cp Glycerin (10 ppg)
100 cp Glycerin (11.5 ppg)
Fig.10. Proposed Correlation of ZNLF Holdup (Viscosity Effect)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 2 4 6 8 10 12 14 16 18
V
SG
(ft/sec)
H
L
0
100 psig 200 psig
300 psig 400 psig
500 psig 1000 psig
Fig.11. Proposed Correlation of ZNLF Holdup (Pressure Effect)
Falling Film
L
TB
, Length of
Taylor Bubble
L
SU
, Length of
a slug unit
SG C
v C v
0

gD Fr v C v
SG C
5 . 0
0
+
ZNLF: H
GTB
=H
GLS
Non-Zero Net
Stable Slug Flow
L
LS
, Length of
Liquid Slug
L

H
GLS
V
GLS
V
LLS
H
GTB
Viscous Zero Net
Unstable Slug Flow
Non-viscous Zero Net
Unstable Slug Flow
Fig.12. Schematic of ZNLF Regimes in Comparison with Non-
ZNLF
INPUT DATA
Casing I.D 4 in
Tubing O.D 2 in
Depth to perforation zone 10000 ft
Cullender-Smith Depth to annulus fluid level 3000 ft
Method Producing casing head pressure 200 psig
Producing casing head temperature 60 F
Casing head gas rate 200 Mscfd
Gas specific gravity 0.8
Pressure Traverse Liquid specific gravity at 60 F 0.952
Calculation Surface tension at standard condition 60 dynes/cm
Geothermal gradient 0.010625 F/ft
Total Produced GOR 700 scf/STB
Number of Cullender-Smith Increments 200
Number of Pressure Traverse Increments 300
ZNLF
Perforation Zone
Liquid
Gas
?
OUTPUT R U N MAIN
Fig.13. Input Slide
ESTIMATION OF BOTTOMHOLE PRESSURE IN PUMPING OIL WELLS:
SPE 63047 EFFECT OF HIGH-VISCOSITY FLUIDS AND CASINGHEAD PRESSURE 9
OUTPUT DATA
200 psig BHFP at 10000 ft 1911.01 psig
200 Mscfd Flowing pressure at 3000 ft 218.66 psig
Temperature at 10000 ft 186.25 F
Temperature at 3000 ft 111.88 F
Viscosity at 10000 ft 11.15 cp
Viscosity at 3000 ft 76.06 cp
218.6631 psig Dead-Oil Viscosity at 60 F 463.14 cp
3000 ft Bubble-point Pressure at 10000 ft 5993 psig
1911.0138 psig
10000 ft
INPUT MAIN GRAPH DATA
by Hyoung - Jin An
hantiger@hotmail.com
Fig.14. Output Slide
3000
4000
5000
6000
7000
8000
9000
10000
0 500 1000 1500 2000
Pressure, Psig Pressure, Psig Pressure, Psig Pressure, Psig
D
e
p
t
h
,
f
t
D
e
p
t
h
,
f
t
D
e
p
t
h
,
f
t
D
e
p
t
h
,
f
t
Fig.15. Pressure Profile
Liquid Gas Gradient Mixture
Depth, Pressure, Temp, Viscosity, Density, Density, Rs, BO, Bt, CO, VSG, correction gradient, z-Factor
ft psig F cp lb/ft
3
lb/ft
3
scf/STB res bbl/STB res bbl/STB psi
-1
ft/sec factor psi/ft
3000.00 218.66
112.00 75.77 57.65 0.90 30.34 1.036 10.035 1.7663E-05 2.388 0.3719 0.1530 0.9832
3023.33 222.23
112.25 75.20 57.64 0.92 30.79 1.036 9.893 1.7557E-05 2.353 0.3738 0.1537 0.983
3046.67 225.82
112.49 74.64 57.63 0.93 31.24 1.037 9.755 1.7452E-05 2.318 0.3756 0.1545 0.9827
3070.00 229.43
112.74 74.08 57.61 0.94 31.70 1.037 9.620 1.7349E-05 2.284 0.3775 0.1553 0.9825
3093.33 233.05
112.99 73.53 57.60 0.96 32.15 1.037 9.488 1.7249E-05 2.252 0.3794 0.1560 0.9822
3116.67 236.69
113.24 72.98 57.59 0.97 32.61 1.038 9.360 1.7149E-05 2.220 0.3812 0.1568 0.982
3140.00 240.35
113.49 72.44 57.58 0.99 33.07 1.038 9.234 1.7052E-05 2.188 0.3831 0.1576 0.9818
3163.33 244.02
113.73 71.90 57.56 1.00 33.53 1.038 9.112 1.6956E-05 2.158 0.3850 0.1583 0.9815
3186.67 247.72
113.98 71.37 57.55 1.01 34.00 1.039 8.992 1.6862E-05 2.128 0.3869 0.1591 0.9813
3210.00 251.43
114.23 70.84 57.54 1.03 34.46 1.039 8.875 1.6769E-05 2.099 0.3887 0.1598 0.981
3233.33 255.16
114.48 70.31 57.52 1.04 34.93 1.039 8.761 1.6678E-05 2.070 0.3906 0.1606 0.9808
3256.67 258.91
114.73 69.79 57.51 1.06 35.40 1.040 8.650 1.6589E-05 2.042 0.3925 0.1614 0.9805
3280.00 262.67
114.97 69.27 57.50 1.07 35.87 1.040 8.541 1.6500E-05 2.015 0.3944 0.1621 0.9803
3303.33 266.46
Fig.16. Data Sheet