review of methods and correlations for the analysis · pdf filereview of methods and...
TRANSCRIPT
Gas Production and Handling Advances
REVIEW OF METHODS AND CORRELATIONS FOR THE ANALYSIS OF TRANSPORT LINES WITH MULTIPHASE FLOW
Authors: Ruth Anselmi, Alberto J. Baumeister, Katiuska C. Márquez
INELECTRA, S.A.C.A.
ASBTRACT
Multiphase flow transport lines sizing requires different or additional criteria to
those used for the design of single-phase flow lines or two-phase flow lines.
Unlike lines with a single phase, an oversized multiphase flow line can generate
serious operational problems.
Flow assurance is essential to the transportation of multiphase fluids, and
requires evaluations in steady state and dynamic simulations in order to analyze
the potential problems that could arise when changes occur in the flow regime
throughout the studied system.
Flow assurance consists of the analysis (hydraulic, thermal and thermodynamic)
required to maintain uninterrupted production from the reservoir up to the
reception facilities with minimum investment and operation costs.
During the last decades, several hydraulic models have been generated for
predicting and classifying flow patterns present in multiphase flow. In most cases,
these classifications have been developed for flow in horizontal and vertical
pipelines independently.
The limitations of these empirical correlations gave way to the mechanistic
models that originated during the mid-1970's. These models try to describe the
phenomenon of two-phase and multiphase flow using balance sheets and
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 1
Gas Production and Handling Advances
equations that describe the process, and therefore are not limited to specific
conditions such as type of fluid or pipe diameters.
Through this paper a comparison between these models is developed, and the
procedure and criteria currently used for sizing multiphase flow lines is described.
Advantages of the use of dynamic simulators which integrate new mechanistic
models for such evaluation are also discussed.
1. INTRODUCTION
The simultaneous flow of gas and liquid occurs frequently in the oil and chemical
industry, particularly in pipeline transportation, chemical reactors and heat
exchangers. For this reason, the study of the characteristics and mechanisms of
the two-phase flow, has attracted great interest since the decade of the 1940’s,
especially in the oil industry, where the possibility of using a single line for the gas
and oil transportation from the field to the processing plant, results in a substantial
reduction in costs. The design of these pipes requires an estimation of the
multiphase pressure drop as precisely as possible.
Figure 1 shows examples of multiphase flows in the production of hydrocarbons. The
combination of fluids can be present from the reservoir to the reception facilities. In
some cases, some of these phases can be originated during transport, mainly due to
the decrease in temperature, eg. the formation of hydrocarbon condensate, free
water, paraffin or hydrate [Azócar, 2007].
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 2
Gas Production and Handling Advances
Multiphase Flow
Gas + Liquid Drops (Hydrocarbon o Water)
Hydrocarbon liquid + Bubble of gas and/or Water drops
Free water + Bubble of gas and/or Hydrocarbon drops
Hydrates + Another Phase
Paraffins(wax), Asphalthenes o Naphthene + Another
Sand+ another phase
Gas + Hydrocarbon liquid + Water
Figure 1. Multiphase flow during hydrocarbon production [Azócar, 2007].
Gas Phase Drop liquid within Gas Phase
Gas bubble within liquid phase
Liquid hydrocarbon phase
Liquid water Phase
sand Water Drop within hydrocarbon phase
Hidrocarbon drop within wather phase
Figure 2. Example of Multiphase flow present in transport lines [Azócar, 2007].
For the calculation of pressure drop in multiphase flow, it is customary to divide the
total pressure gradient in three components: friction, acceleration and elevation. Each
calculated separately and then combined. Due to the complexity of these calculations
for the two-phase flow, and the need to develop methods for easy application in
industry, the first approaches that were developed over 70 years ago to solve this
problem were the empirical type. The correlations that have been most commonly
used are those developed by Dukler et al. (1964) and Beggs and Brill (1973) for
horizontal and slightly slant flow in pipes, and the Hagedorn and Brown (1965), and
Duns and Ross (1963) for flow into wells (vertical pipes). These approaches were
very successful in solving two-phase flow problems, obtaining with them a maximum
error of ±30%. However, the empirical correlations have never explained why and
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 3
Gas Production and Handling Advances
how the phenomenon of two-phase flow occurs [Gomez et al., 1999]. The other big
disadvantage of these correlations is that its use is limited to conditions similar to
those that were developed; outside them the deviations in the values predicted
increase significantly. Thus, when these correlations are used to predict the pressure
drop in larger diameter pipes, or fluids of different properties (more viscous, for
example), the values obtained vary appreciably from the experimental values [Badie
et al., 1999].
The empirical correlations proposed over the years to calculate pressure drop and
liquid holdup have been based on various approaches. The first of them were based
in so-called homogeneous models, which assumed that the liquid and gas phases
travel at the same speed (no slippage between phases), and make no consideration
of flow patterns. Subsequently, the separated flows models were developed, which
supposed that the liquid and gas phases have different speeds (slipping is taken into
account), but also paid attention to the flow regime. A substantial improvement of the
separated flows models occurred when the flow patterns were considered, and using
the same principles, different models were proposed for each flow regime present in
the system.
Due to limitations of the empirical correlations, a new approach emerged in the mid-
1970’s, called mechanistic models. This approach aims to clarify the physical
phenomenon. The mechanism that causes multiphase flow is determined and
modeled mathematically. The first objective of this approach is to predict the flow
regime in a given system. Then, a model is developed separately for each flow
pattern to predict its hydrodynamics and heat transfer [Gomez et al., 1999].
The mechanistic models developed in the last two decades have been formulated
separately for horizontal and vertical piping. The models for horizontal pipes can also
be used for slightly slant piping (±10 °). The first model, which has lasted the most for
the prediction of flow patterns, is the Taitel and Dukler (1976) model. A
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 4
Gas Production and Handling Advances
comprehensive mechanistic model, is that which can predict flow patterns initially,
and then develop separate models for the prediction of the pressure drop for each of
these patterns. An example of this type was introduced by Xiao et al. (1990) for piping
design [Gomez et al., 1999].
The objective of this paper is to summarize the differences between the empirical
correlations more commonly employed in the industry vs. some of the most recent
mechanistic models that are integrated into some commercial simulators. These
simulators allow not only the prediction of pressure drop and liquid holdup in pipe
sections, but also the performance of transient state studies with the aim of analyzing
the behavior of systems under various disturbances during the same operation.
These analyses allow more appropriate and optimal designs to those made in
previous decades.
2. CONTENT 2.1. Basic Definition The calculation of the pressure gradient requires knowing the velocity and properties
of fluids such as density, viscosity and, in some cases, surface tension. When these
variables are calculated for two-phase flow, the use of certain rules and unique blend
definitions are required [Beggs and Brill, 1991]. The most important properties are the
following:
A) Liquid holdup:
The liquid holdup is defined as the ratio of the amount of liquid volume in a segment
of pipe and the total volume of this segment.
volume segment Pipesegment pipe in volume LiquidHL = (1)
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 5
Gas Production and Handling Advances
The definition of HL varies between zero for gas phase only, and one for liquid phase
completely. The remaining volume of the pipe occupied by gas is called gas holdup or
vacuum fraction.
LG H1−=α (2)
B) No-slip holdup :
The liquid holdup without slippage is the ratio between the volume of the liquid in a
line segment divided by the total volume of this segment, considering that both
phases travel at the same speed. It can be calculated directly with the volume flow
rates.
GL
LL QQ
Q+
=λ (3)
The gas holdup without slippage is defined as:
LG 1 λλ −= (4)
The difference between the holdup with and without slip is a measure of the degree of
slippage between phases.
C) Superficial Velocity:
The superficial velocity of a fluid phase is the velocity that this phase would show if it
flowed alone through the cross section of the line. The superficial velocity is the
volumetric flow rate of one of the phases per unit area.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 6
Gas Production and Handling Advances
G
GSG A
Qv = (5)
L
LSL A
Qv = (6)
The speed of the mixture is the total volumetric flow rate per unit area and is given by:
SGSLT
GLM vv
AQQv +=
+= (7)
D) Actual (in situ) Velocity
The actual (or in situ) velocity is the volumetric flow rate divided by the area occupied
by that phase. Therefore, actual velocity of liquid or gas phase can be calculated
using the following expressions:
L
SL
LT
LL H
vHA
Qv =⋅
= (8)
L
SG
GT
GG H1
vA
Qv−
=⋅
=α
(9)
E) Slip Velocity
The slip velocity is the relative velocity between the two phases, and is defined as the
difference between the actual velocities of the gas and the liquid.
LGS vvv −= (10)
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 7
Gas Production and Handling Advances
2.2. Flow Patterns
One of the most important factors in the study of multiphase flow are the flow
patterns, which are understood as the different configurations that phases acquire in
the pipeline, when both travel simultaneously. There are various classifications based
on visual observations carried out by different authors. The prediction of these
regimes is of great importance in determining the pressure drop and liquid holdup
because much of the empirical correlations and mechanistic models used in these
calculations depends on the flow pattern existing under different working conditions.
The flow patterns usually identified in the industry are defined below, in a general
manner.
Horizontal Flow
Segregated Flow: This flow is presented when liquid phase velocity is low, while the
velocity of the gas phase can go from low to moderate. Within the classification of
segregated flow are: stratified flow, wavy flow and annular flow.
Intermittent Flow: This flow occurs generally when the liquid phase has a moderate
velocity, while the gas phase is between moderate and high. Within the classification
of intermittent flow are: plug flow and slug flow.
Distributed Flow: Within this classification one of the two phases is dominant over
the other, because its flow rate, and thus its velocity, is much higher than the other’s.
In this category of distributed flow is included: bubble flow and mist flow.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 8
Gas Production and Handling Advances
Intermittent Distributed Segregated
Elongatea bubble flow
Slug Flow
Dispersed bubble flow
Mist
Stratified Smooth flow
Stratified Wavy Flow
Annular
Figure 4. Flow Patterns for gas-liquid systems in horizontal pipes, terminology
from Beggs and Brill, adapted from [Beggs and Brill, 1991].
Vertical Flow
Bubble Flow: The gas phase is distributed in the form of bubbles immersed in a
continuous liquid phase.
Bubble - Liquid Slug Flow: As the concentration of bubbles grows by the presence
of a higher quantity of gas, bubbles group or coalesce into one whose diameter
approaches the pipe diameter.
Transition flow, Liquid Slug -Annular: With greater flow rate, the bubbles formed in
the bubble flow collapse, resulting in a sparkling and disorderly flow of gas through
the liquid that is displaced to the wall of the channel.
Annular - Bubble Flow: The flow takes the form of a relatively thick liquid film on the
pipe wall, along with a substantial amount of liquid carried by the gas flowing in the
center of the channel.
Annular flow: The liquid film is formed on the wall of the tube with a central part
formed by gas.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 9
Gas Production and Handling Advances
Annular – Liquid Bubble
Bubble Bubble – Slug flow Annular Mist flow
Transition: Liquid Slug – Annular
Figure 5. Flow Patterns for gas-liquid system in vertical pipes, Beggs and Brill
terminology, adapted from [Beggs and Brill, 1991].
2.3. Flow-pattern map
The first approximation to predict flow patterns has been the empirical approach
based primarily on visual observation. Usually the data has been two-dimension
graphs and these figures identify the transition limits between different flow patterns.
These representations are called flow maps. In most cases, the coordinates are
chosen in an arbitrary manner, without a physical basis. For this reason, each map is
useful only at intervals of conditions similar to those in which the data was acquired,
and extending them to other conditions is uncertain. Figure 6 shows one of the first
flow maps developed and widely used in the industry as is the Beggs and Brill map.
Commonly, parameters such as surface and mixing velocities, among others, are
used as coordinates of these maps, because these parameters are the most
characteristic for multiphase flow and the most frequently used in the analysis of this
phenomenon. Other authors, in attempts to extend the validity of their maps have
chosen dimensionless coordinates or correction factors for the physical properties
[Shoham, 1998].
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 10
Gas Production and Handling Advances
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 11
Figure 6. Flow Map, Beggs and Brill (1973) [Brill and Beggs, 1991].
2.4. Models and Correlations for Total Pressure Drop Calculation Like the single-phase case, the total pressure gradient is usually divided into three
components which are derived through an energy balance. An elevation component
(subscript e) which represents the change of potential energy or elevation on the line,
a frictional losses component (subscript f) and an acceleration component (subscript
a) which represents the change of kinetic energy.
afe dLdP
dLdP
dLdP
dLdP
⎟⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛+⎟
⎠⎞
⎜⎝⎛= (11)
Many correlations have been developed to predict pressure drop, which differed on
how to calculate these three components. The definitions of each term for single-
phase flow are usually adapted for two-phase flow assuming that the gas-liquid
mixture is homogeneous for a finite volume of the line.
Contents of liquid at the entrance, λL
Fr
Gas Production and Handling Advances
Table 1. Summary of the effects of frictional, elevation and acceleration terms
according to flow type.
Flow Type Friction term Elevation term Acceleration term
Single Phase
Friction against the surface of pipe
Independent from flow Liquid: Dependent on elevation difference Gas: Negligible, except for very high-pressure systems
Generally negligible, but applies for high speed depressurizations
Multiphase flow
Friction of the phases against the surface of the line. Friction between the phases
Density of the phases. Angle of inclination. Liquid Amount (flow variations).
Generally negligible, but applies for high speed depressurizations
In general, the frictional component differs between one model and another mainly in
the way the friction factor is determined and the variety of flow patterns. Many
correlations try to relate the friction factor with different definitions of the Reynolds
number. On the other hand, the acceleration component is completely ignored by
some authors and ignored in some flow patterns by others.
Steady state process simulators have been traditionally used in the industry to
calculate and predict the behavior of two-phase mixtures; these have extensive
databases to estimate physical properties and use empirical correlations to predict
pressure drop and flow regime in each pipe section. Among these simulators is
PIPEPHASE®.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 12
Gas Production and Handling Advances
2.5. Flow Assurance
The objective of flow assurance is to guarantee or maintain uninterrupted production
from the reservoir up to the reception facilities with minimum investment and
operation costs.
Flow assurance is now one of the main topics for the design of a system for
hydrocarbons production and the specification of production processes.
The topics considered in flow assurance analysis are evaluated at steady and
transient state. The analysis includes the evaluation of risks and uncertainties
associated with the operational procedures, and contribute to a better economic
estimate of production facilities.
2.6. Simulation for non-stationary multiphase flow In recent years there has been diffused use of simulators that integrate mechanistic
models for the estimation of pressure drop. Using these simulators, it is possible to
size pipes considering flow assurance. Estimations made with these simulators
achieve a better reproduction of transportation systems behavior because they allow
transient-state simulation. Among these simulators are OLGA® from Scandpower
Petroleum Technology, and PIPEPHASE® with the TACITE® module from the
Institute Français du Petrole.
The mechanistic model called OLGA is based on separate balances for mass
conservation of the gas phase and the liquid phase, as well as the drag of liquid
droplets. The formulation of this model includes additional equations for momentum
balances for each phase, and combined mixture energy balance.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 13
Gas Production and Handling Advances
For this model, like any other mechanistic model, the implementation of conservation
equations is dependent on the flow pattern. In particular, friction factor and wet
perimeter terms are based on the prediction of the geometry of the flow distribution.
Therefore, the first step of the algorithm used in the model, is the determination of the
flow pattern based on the local distribution of the two-phase flow parameters.
OLGA has been compared with data from various experimental facilities, covering a
wide range of pipe diameters, fluids, angles of inclination and operating conditions.
Most of the information was obtained from experiments of SINTEF Two-Phase Flow
Laboratory in Norway. The model has also been tested successfully in a wide range
of oil production facilities.
Similarly, TACITE is a model that rigorously resolves mass balances for each
component of the mixture, as well as mass and energy balances for the multiphase
mixture. The mechanistic model depends on the flow pattern and has been
extensively validated with experimental field data. The model enables the tracking of
variations in the fluid composition and behavior of the phases along the pipeline for
changes in the inlet conditions and elevation during non-stationary state.
Table 2 shows a comparison between TACITE ® and OLGA ® simulators, which
have been used in recent projects developed by the company INELECTRA SACA.
The main advantage of the program OLGA ® from Scandpower Petroleum
Technology is that, according to project requirements, each module for flow
assurance can be purchased separately. However, the purpose of this comparison is
not to determine which of the models incorporated in these simulators is more
accurate for sizing multiphase pipes, but to show the versatility of each one. The
selection of the simulator to be used in each case depends on customer
requirements, study case characteristics, etc.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 14
Gas Production and Handling Advances
Table 2. Comparison between commercial simulators
Basic Evaluation TACITE® OLGA®
Mechanistic Model
Individual Pipes
Recollection Networks ---
Transient State Simulations
Slug Flow Formation Analysis
Control System for Slug Catcher Design.
Possibility to view different variables in profiles or time graphs. ---
Flow Assurance Modules
Three Phases considering separation of water and hydrocarbon, and velocity difference between liquid phases (slip)
---
Liquid Slugs tracking ---
MEG tracking as inhibitor of hydrates formation ---
Hydrates Formation ---
Paraffin Deposition ---
CO2 Corrosion ---
Bundled Lines ---
Heat transfer tridimensional for buried lines ---
For non stationary state simulations it is required first to conduct a sensitivity analysis
at steady state to define the preliminary limits of operation. With this analysis it is
possible to establish the diameter of the pipe and peak flow through it taking into
account the maximum permitted backpressure on the system and verifying that the
fluid velocities do not exceed the established design criteria. Similarly, it is possible to
establish the minimum system flow rate in order to avoid accumulations of liquid
beyond the capability of the reception facilities. This analysis can also be used to
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 15
Gas Production and Handling Advances
study the temperature profile along the pipe to determine the flow ranges where
hydrates formation and / or deposition of paraffins could take place.
Subsequently, the development of the sensitivity analyses in non stationary state,
which are indispensable for flow assurance of multiphase flow lines, is required. The
presence of at least two phases causes changes as a function of time for variables
such as pressure, temperature, holdup and flow pattern.
The simulations in transient state allow establishing the minimum and maximum
values of the different studied variables (such as pressure drop, fluid, etc.) which
occur under changing operating conditions (eg Start/Stop conditions, increase or
decrease of load, changes in topography, etc.). The visualization of the system
behavior under these circumstances allows optimizing the design and establishment
of the operation ranges.
2.7. Results Comparison between Empirical Correlations vs. Mechanistic Models Following are the results of the evaluation of a section of pipeline that transports
natural gas and its condensates, using the Beggs - Brill - Moody (BBM) correlation,
and OLGAS (OLGAS is the steady-state version of OLGA from ScandPower AS, and
is an additional module that can integrate with PIPEPHASE®).
The purpose of this exercise was to demonstrate the difference in the values obtained
for pressure drop, friction factor, and liquid holdup at steady state for the same gas
flow (100 MMSCFD) and same operating conditions (Pressure defined at exit point:
500 psig, defined temperature at entry point: 150° F). In order to determine the
optimum diameter of the pipe, a sensitivity study was conducted in a range between
10" and 14" (considering carbon steel pipe, wall thickness: 0.5 in., buried pipe, soil
and coating properties known). One vertical branch (length: 1000 feet upstream, 1000
ft downstream) and one horizontal branch (length: 2000 ft) were considered.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 16
Gas Production and Handling Advances
In the case of a horizontal pipe, there is a significant difference between both models
for diameters of 10" and 12". The main reason for this difference is based on the
calculated values for the friction factor; in the case of OLGAS, the friction factor is
nearly constant for all diameters evaluated, while BBM presents a jump between 10"
and 12" (see Figure 7).
0
5
10
15
20
25
30
35
40
45
50
10 12 14 16
Pressure Drop [psi]
Internal D iam eter [ in]
B B M
O L G A S
Figure 7. Total pressure drop vs. internal diameter in horizontal pipe
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 17
Gas Production and Handling Advances
0.00
0.05
0.10
0.15
0.20
0.25
0.30
10 12 14 16
Friction Factor
Internal Diameter [in]
BBM
OLGAS
Figure 8. Friction factor vs. internal diameter in horizontal pipe.
0.00
0.01
0.01
0.02
0.02
0.03
10 12 14 16
Liquid Holdoup
Internal Diameter [in]
BBM
OLGAS
Figure 9. Liquid Holdup (slip) vs. internal diameter in horizontal pipe.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 18
Gas Production and Handling Advances
-0.03
-0.02
-0.02
-0.01
-0.01
0.00
10 12 14 16Frictional Pressure Drop [psi/ft]
Internal Diameter [in]
BBM
OLGAS
Figure 10. Frictional Pressure Drop vs. internal diameter in horizontal pipe
In the case of vertical upward pipe model, BBM predicts large pressure drops from an
internal diameter of 14". Among the variables that could cause the results shown is
the slip liquid holdup and its effect on the pressure gradient per elevation. In the case
of BBM, the slip liquid holdup increase can be observed, as well as the elevation
gradient. Meanwhile, OLGAS predicts a behavior almost constant for these two
variables.
With respect to the discontinuance of the total pressure drop in BBM (Figure 11),
clearly visible on the 12" diameter, this is a compensation of the frictional and
elevation pressure drops, see Figure 15.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 19
Gas Production and Handling Advances
10
15
20
25
30
35
40
45
10 12 14 16 18 20 22
Pressure Drop [psi]
Internal Diameter [in]
BBM
OLGAS
Figure 11. Total pressure drop vs. internal diameter in vertical pipe upward flow
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
10 12 14 16 18 20 22
Liquid Holdup
Internal Diameter[in]
BBM
OLGAS
Figure 12. Slip Liquid Holdup vs. internal diameter in vertical pipe upward flow
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 20
Gas Production and Handling Advances
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.005
10 12 14 16 18 20 22
Frictional Pressure Drop [psi/ft]
Internal Diameter [in]
BBM
OLGAS
Figure 13. Frictional pressure drop vs. internal diameter in vertical pipe upward flow
-0.045
-0.040
-0.035
-0.030
-0.025
-0.020
-0.015
-0.010
10 12 14 16 18 20 22
Elevation Pressure Drop[psi/ft]
Internal Daimeter [in]
BBM
OLGAS
Figure 14 Elevation pressure drop vs. internal diameter in vertical pipe upward flow
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 21
Gas Production and Handling Advances
-0.045
-0.040
-0.035
-0.030
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
10 12 14 16 18 20 22
Pressure Drop [psi/ft]
Internal Diameter [in]
Frictional Pressure Drop
Elevation Pressure Drop
Total Pressure Drop
Figure 15. Total pressure drop according to BBM vs. internal diameter in vertical pipe upward flow
In the case of the vertical pipe for downward flow, the biggest difference in the result
of total pressure drop is presented in diameters of 10" and 12", but after 14" the
difference between the two models is lower. For the BBM model, a discontinuity in the
values of liquid holdup can be observed for diameters of 10" and 12", affecting the
elevation pressure drop as well.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 22
Gas Production and Handling Advances
-15
-10
-5
0
5
10
15
20
10 12 14 16 18 20 22
Pressure Drop [psi]
Internal Diameter [in]
BBM
OLGAS
Figure 16. Total pressure drop vs. internal diameter in vertical pipe downward flow
0.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
10 12 14 16 18 20 22
Liquid Holdup
Internal Diameter [in]
BBM
OLGAS
Figure 17. Slip Liquid Holdup vs. internal diameter in vertical pipe downward flow
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 23
Gas Production and Handling Advances
-0.035
-0.030
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
10 12 14 16 18 20 22
Fictional Pressure Drop [psi/ft]
Internal Diameter [in]
BBM
OLGAS
Figure 18. Frictional pressure drop vs. internal diameter in vertical pipe downward flow
0.011
0.012
0.012
0.013
0.013
0.014
0.014
10 12 14 16 18 20 22
Elevation Pressure Drop [psi/ft]
Internal Diameter [in]
BBM
OLGAS
Figure 19. Elevation pressure drop vs. internal diameter in vertical pipe downward flow
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 24
Gas Production and Handling Advances
3. CONCLUSIONS
• A comparison of the results for the evaluated example shows that considerable
differences may occur in the pressure drop calculation using empirical
correlations and mechanistic models.
• Most of the curves obtained with BBM have “jumps” between one diameter and
another. The reasons for this may be:
a) Empirical correlations used for pressure drop estimation are limited to
the range of data to which they are based on. This fact reduces their
reliability for the fluid types and conditions that can be found in
production and transportation facilities.
b) Many models have discontinuities in determining the transitions
between the flow patterns.
• The mechanistic models are, according to the literature, more reliable in a wide
range of fluids, process variables, pipe inclinations, etc. Additionally, they
associate pressure drop calculations with liquid holdup and flow patterns,
which ensure continuity in optimal results.
4. REFERENCES
• Azócar, A. Dimensionamiento de Líneas de Transporte con Flujo
Multifásico. INELECTRA S.A.C.A. Process Department Work Instructions.
• Brill, J., and Beggs, H. Two-Phase Flow in Pipes. Sixth Edition (without
editorial name). 1991.
• Gómez, L.E., O. Shoham and Z. Schmidt. “A Unified Mechanistic Model for
Steady-State Two Phase-Flow in Wellbores and Pipelines”. SPE International.
October, 1999.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 25
Gas Production and Handling Advances
• Badie S., C.P. Hale, C.J. Lawrence and G.F. Hewitt. “Pressure Gradient And
Holdup In Horizontal Two-Phase Gas-Liquid Flows With Low Liquid Loading”.
International Journal of Multiphase Flow, vol. 26, 1999.
XVIII Gas Convention, AVPG, Caracas, Venezuela, May 27 - 29 th, 2008 Page 26