Escolar Documentos
Profissional Documentos
Cultura Documentos
Eissa M. Al-Safran
Kuwait University
James P. Brill
University of Tulsa
All rights reserved. No portion of this book may be reproduced in any form or by any means, including electronic
storage and retrieval systems, except by explicit, prior written permission of the publisher except for brief pas-
sages excerpted for review and critical purposes.
Disclaimer
This book was prepared by members of the Society of Petroleum Engineers and their well-qualified colleagues
from material published in the recognized technical literature and from their own individual experience and
expertise. While the material presented is believed to be based on sound technical knowledge, neither the Society
of Petroleum Engineers nor any of the authors or editors herein provide a warranty either expressed or implied
in its application. Correspondingly, the discussion of materials, methods, or techniques that may be covered by
patents implies no freedom to use such materials, methods, or techniques without permission through appropriate
licensing. Nothing described within this book should be construed to lessen the need to apply sound engineering
judgment nor to carefully apply accepted engineering practices in the design, implementation, or application of
the techniques described herein.
ISBN 978-1-61399-492-4
http://www.spe.org/store
service@spe.org
1.972.952.9393
Central Idea
The central idea of this textbook is to convey the process of designing and operating the components of a
petroleum production system that involve multiphase flow, as shown in the inner circle of Fig. 1. This requires
determining the multiphase flow pattern, liquid volume (holdup), and pressure gradient along the piping system,
as shown in the middle circle of Fig. 1. The input parameters to determine these three variables are the in-situ fluid
physical properties and flow rates. This textbook takes the reader from the outer circle inward in a step-by-step
and simple approach, providing the conceptual understanding and mathematical tools to carry out a design and/
or operational study.
Fluid Properties
Flow Pattern
Piping System
Design and
Operation
Liquid Pressure
Volume Gradient
vii
Organization
Predicting flow behavior when multiphase flow occurs in wells and pipelines requires an understanding of con-
cepts that are not part of the curriculum in most engineering disciplines. Before multiphase flow technology can
be mastered, one must first have an adequate knowledge of single-phase flow fluid mechanics, vapor/liquid equi-
librium (VLE), and fluid physical properties for multicomponent hydrocarbon systems. Thus, a significant part of
this textbook will be devoted to a review of these important topics.
Chapter 1 starts with the basic definition of multiphase flow, followed by a brief description of the significance
of multiphase flow during the production and transportation of oil and gas. The overall production system is pre-
sented, together with concepts that must be considered by a production engineer. The future outlook of multiphase
flow technology is then presented. A brief review of the history of multiphase flow is presented in which some
landmark publications are identified that had a lasting influence on our understanding of the total system involved
in flowing fluids from the reservoir to surface storage and processing facilities.
Chapter 2 concentrates on a review of important single-phase, steady-state flow concepts for both incompress-
ible Newtonian and compressible fluids. A description of single-phase non-Newtonian fluid flow is also presented.
Chapter 3 presents homogeneous two-phase flow in pipes. It starts with a physical description, followed by
calculation of homogeneous mixture properties and pressure gradient. The authors present several applications of
homogeneous two-phase flow with practical examples. A brief introduction of the drift flux model is presented at
the end of Chapter 3.
Chapters 4 and 5 are the main chapters of this textbook and present nonhomogeneous, two-phase flow in pipe-
lines and wells, respectively. These chapters start with physical descriptions of the slippage phenomenon and flow
pattern, which are followed by flow pattern, liquid holdup, and pressure gradient predictions using both empirical
and mechanistic modeling approaches.
In Chapter 6, the authors present the physical and theoretical aspects of two-phase flow through restrictions
and piping components for both homogeneous and nonhomogeneous flows. The chapter begins with the p hysical
concepts of single- and two-phase flow through chokes, followed by flow-regime-dependent empirical and
theoretical predictions. Pressure drop and flow rate across chokes for homogeneous and nonhomogeneous flow
are then discussed and shown by examples.
Chapter 7 presents a simple approach for single-phase and homogeneous two-phase heat transfer in a produc-
tion system. This chapter is considered introductory material for multiphase flow assurance.
Chapter 8 is a detailed presentation of flow assurance that includes the definition, significance, types, causes,
predictions, and remedies of the phenomena. Chapter 8 concludes with a presentation of the integrated flow
assurance work flow process, which summarizes the entire chapter. Several examples are presented throughout
the chapter to demonstrate the solution procedures of the presented models.
Chapter 9 is dedicated to the Tulsa University Fluid Flow Projects’ unified mechanistic model. The c hapter
highlights the differences between the comprehensive mechanistic models presented in Chapters 4 and 5, followed
by the theoretical features of the unified model. The calculation procedure of the unified model is demonstrated
in three long-hand solved examples of horizontal and vertical flows.
Chapter 10 combines all material covered in previous chapters into a chapter on the overall production system.
The chapter concludes with a comprehensive design example problem.
Chapter 11 is an introduction to transient multiphase flow in pipes. In this chapter, the concepts and application
of transient flow are presented. A comparison between steady-state and transient flows is also presented, and a
guidance is provided on when to use a steady-state solution to approximate a transient event. The chapter ends
with a simplified transient modeling approach and a description of two commercial, transient, multiphase flow
simulators used in the oil and gas industry.
Appendix A lists the nomenclature used in this book. Appendices B and C contain methods to predict fluid
physical properties and in-situ volumetric flow rates using black-oil and compositional models, respectively.
Appendices B and C contain example problems to demonstrate the solution procedures.
ix
1.2.1 Wells. As the pressure drops in production tubing because of fluid flow and reduction of hydrostatic head,
dissolved gas in the oil evolves, forming a gas/liquid two-phase flow. In addition to oil and gas phases, produced
water and solids, such as sand, introduce multiphase flow in the well tubing, which further complicates the flow
behavior and characteristics. Furthermore, as the temperature drops going up the well, dissolved solids may pre-
cipitate, adding a solid phase to the flow stream, which may be either deposited within or transported along the
well tubing. In the case of gas lift wells, a gas phase is injected into the tubing, causing two-phase flow. Horizontal
well production exhibits a complex, undulated multiphase flow with complex fluid influx geometry, making mul-
tiphase flow even more challenging.
1.2.2 Pipelines. In offshore fields, produced fluids are transported to shore through long, large-diameter export
pipelines for subsequent separation and processing. Tiebacks are often long pipelines that transport untreated
produced fluids from marginal fields to an existing platform. These pipeline systems almost always exhibit mul-
tiphase flow. Onshore flowlines and export pipelines are multiphase flow systems that require proper design and
operation for safe and economic business. Accurate prediction of flow pattern, liquid holdup, pressure drop,
and flow characteristics along these pipeline systems is essential to enhance not only pipeline design, sizing,
and routing, but also design of downstream separation and processing facilities to ensure safe and economic
business operation (see Example 1.1).
1.2.3 Wellhead Chokes. Wellhead chokes and other restrictions are integral parts of an oil and gas production
system. Multiphase flow across these restrictions is common and complex. The diversity of the flow through
restrictions can cause critical or subcritical, as well as homogeneous or nonhomogeneous, multiphase flow. Pre-
dicting the flow regime (i.e., critical vs. subcritical) and the pressure drop is crucial for estimating and controlling
the system flow rate. In addition, restrictions can impose flow assurance problems, among them formation of
emulsions, deposition of solids as a result of sudden cooling, and erosion of pipe caused by high velocities during
flow through restrictions.
1.2.4 Other System Components. Recent in-line technologies such as multiphase flowmeters, multiphase flow
pumps, water knockout systems, and gas/liquid compact cyclones (GLCCs) have become important parts of
the production system, especially in mature fields. Predicting multiphase flow characteristics upstream of these
devices is critical for their efficiency and integrity. In addition, understanding the flow behavior inside internal
parts improves future design and reliability.
Gas
Foam
Oil
Emulsion
Water
Wellbore
Reservoir
pwf pr pe
in a gas/liquid mixture entering the well. Retrograde condensation can result in hydrocarbon liquids condensing
in a gas/condensate reservoir so that a gas/liquid mixture again enters the wellbore. Even when single-phase gas
or liquid flow exists near the bottom of a well, multiphase flow can occur throughout most of the wellbore. This
is because of the evolution of gas from oil or the condensation of gas with reduction of pressure and temperature
as the fluids flow up the well.
Although many of the wells drilled on land tend to be nearly vertical, current trends are toward pad drilling
and horizontal wells. Wells drilled offshore and in other hostile environments such as the Arctic are normally
directional or deviated. Inclination angles can vary from vertical near the surface to horizontal near the production
zone. Flow rates of gas, oil, and water vary widely. Tubing diameters can be as small as 0.0254 m (1 in.) or as
large as 0.2286 m (9 in.). Flow can also occur in a casing/tubing annulus. Depths can range from a few hundred
meters to more than 6,000 m (≈ 20,000 ft). Pressures can be as small as a few hundred kPa or as high as 150 MPa
(≈22,000 psia). Temperatures can be greater than 200°C (≈400°F) or approach the freezing point of water. Oil
viscosities in wellbores can be less than 0.001 Pa·s (1 cp) or greater than 10 Pa·s (10,000 cp).
Fluids entering the wellbore often flow through a complicated well-completion region consisting of perfora-
tions, fractures, and gravel packs. The effect of this region must be included when coupling the well to the res-
ervoir through inflow performance relationship (IPR) procedures. Most wells contain some type of well-control
device that requires produced fluids to flow through a restriction. This can vary from a bottomhole choke to a
remotely controlled, variable-size surface choke. Wells can be produced by artificial-lift mechanisms involving a
submersible pump or gas injection.
The broad variations in flow variables encountered in producing wells has made the development of predic-
tion methods much more difficult. Techniques that work for gas/condensate wells do not necessarily work for oil
wells. Assumptions that are valid for some wells are totally invalid for others.
14
Pipeline Diam. = 91.4 cm
13
11
10
6
0 10 20 30 40 50 60
Gas Flow Rate (106 std m3/d)
7500
Pipeline Diam. = 91.4 cm
7000
Pipeline Diam. = 96.5 cm
6000
5500
5000
4500
4000
3500
0 10 20 30 40 50 60
Gas Flow Rate (106 std m3/d)
The unusual behavior of increasing liquid volume with increasing gas flow rate is a result of retrograde c ondensation
taking place in the pipeline.
Another important design aspect is flow assurance. For example, it is important to accurately predict flow pat-
tern along the pipeline because it is related to pipeline corrosion. A corrosion inhibitor is often injected at the
platform. The inhibitor is transported in the liquid phase along the pipe to protect the pipeline internal wall from
corrosion. It may be very important for efficient operation to design the pipeline so that the entire pipeline wall is
exposed to the corrosion inhibitor. This can be achieved when the flow pattern is a favorable one where the liquid
phase wets the entire internal diameter (i.e., annular or slug flow and not stratified flow). Failure to consider the
flow pattern in pipeline design may lead to rapid pipeline corrosion.
The historical background of attempts to improve the prediction of vertical-lift performance for wells has been
especially interesting. Brill and Arirachakaran (1992) divided this history into three partially overlapping periods.
Fig. 1.4 illustrates the three periods in addition to future projections.
1.5.1 The Empirical Period (1950–75). Most early investigators used two-phase flow data obtained from labora-
tory test facilities, with a few researchers using field data. Fluids were treated as homogeneous mixtures. How-
ever, gas and liquid phases were permitted to travel at different velocities, with slippage effects being accounted
for through empirical liquid holdup correlations. Empirical flow pattern maps such as the ones developed by
Baker (1954), Ros (1961), and Beggs and Brill (1973) were used, often based on dimensionless groups. Steady-
state pressure-gradient equations were developed on the basis of conservation of momentum and mass principles
applied to the homogeneous mixtures. Frictional pressure losses relied on single-phase flow equations, resulting
in extensive use of mixture Reynolds numbers, as in the Beggs and Brill (1973) two-phase flow correlation.
In general, the empirical period resulted in a collection of empirical correlations in which accuracy was limited
by the lack of inclusion of basic physical mechanisms. Even when these mechanisms were partially included,
their investigation was severely hampered by the unavailability of sufficiently accurate instrumentation and real-
time data-acquisition systems.
1.5.2 The Awakening Years (1970–85). The empirical correlations for predicting pressure gradient, coupled
with the introduction of the personal computer (PC) in the early 1980s, dramatically improved practical tools
available to petroleum engineers. Procedures for connecting wells to reservoirs through simple IPR techniques
abounded. The true concept of NODAL™ or production system analysis was born (Brown 1980).
Unfortunately, it was quickly recognized that there were many problems with the methods available. Empiri-
cal flow pattern maps were inadequate. Flow pattern transitions, previously thought to be dependent mostly on
flow rates (or superficial velocities), were found to be very sensitive to other parameters, especially inclination
angle. Empirical liquid holdup correlations for each flow pattern were equally inadequate. The assumption of a
homogeneous mixture was oversimplified. It became clear that no matter how much data were gathered either
Mechanistic models:
Dukler and Hubbard (1975)
Flow pattern map: Taitel and Dukler (1976)
Baker (1954) NODALTM (Brown 1980)
Ros (1961)
Pressure gradient:
Beggs and Brill (1973)
73)
Computational capability
in laboratory test facilities or from carefully tested field installations, the accuracy of the predictions could not
improve without the introduction of more basic physical mechanisms.
Fortunately, progress in this area had already been made by other industries, particularly the nuclear industry,
several years before. Although the fluids used for these studies (steam/water) were trivial by comparison to those
encountered in the petroleum industry, the methods used to formulate conservation equations were much more
advanced. Therefore, the 1970s saw a trend in the petroleum industry to adopt some basic physical mechanisms
already in use in the nuclear industry. Two classic papers dealing with multiphase flow in horizontal pipes by
Dukler and Hubbard (1975) and Taitel and Dukler (1976) clearly show that mechanistic models for slug flow and
flow pattern prediction had already become available.
1.5.3 The Modeling Era (1980–Present). Petroleum industry challenges in the 1980s required a much bet-
ter understanding of multiphase flow technology, beginning the modeling period. Investigators recognized that
improved understanding of multiphase flow in pipes required a combined experimental and theoretical approach.
Sophisticated test facilities were constructed that used new instrumentation for measuring important variables, and
high-speed PC-based data-acquisition hardware and software. This advancement was transformed into improved
mechanistic models to better describe the physical phenomena occurring.
An important improvement in steady-state mechanistic models was the work on predicting flow pattern tran-
sitions for all inclination angles by Taitel and Dukler (1976), Taitel et al. (1980), Barnea et al. (1982a, 1982b,
and 1985), and Barnea (1986, 1987). This opened the door for designing improved models for each of the flow
patterns and linking the various models together through unified flow pattern transition criteria. Combined or
“comprehensive” mechanistic models were published by Ozon et al. (1987), Hasan and Kabir (1988), Xiao et al.
(1990), Ansari et al. (1994), and Chokshi (1994). Their attempts to evaluate the models with field data confirm
that the modeling approach is more accurate and precise than empirical correlations. Furthermore, it is now pos-
sible to continue improvement of these mechanistic models as experimental research is conducted on the basic
mechanisms of multiphase flow.
At the same time that improved experimental research was being conducted, efforts were expanded in devel-
oping improved theoretical methods. The two-fluid modeling approach pioneered by the nuclear industry was
adopted for the development of transient codes for application to petroleum industry problems by Taitel et al.
(1980), Black et al. (1990), Bendiksen et al. (1991), and Pauchon et al. (1993). This approach involves writing
separate equations for each phase that describe conservation of mass, momentum, and energy, resulting in a
six-equation problem that must be solved simultaneously using numerical simulation techniques. Empirical
correlations and simplified closure relationships were still necessary for some parameters. Improved correla-
tions for these parameters became possible as a result of the experimental research being conducted. The result-
ing transient codes are capable of simulating a variety of applications that are time-dependent, such as pipeline
inlet/outlet flow rates or pressure changes, pipeline pigging, startup/shutdown, and terrain slugging.
A recent improvement in the development of steady-state mechanistic modeling is the “unified” modeling
of multiphase pipe flows (Zhang et al. 2003b). Unified modeling is a new approach in which the predictions
of both flow pattern transition and flow behavior are incorporated into a single model based on slug dynam-
ics. This modeling, which covers the entire inclination-angle range (−90° to +90° from horizontal), eliminates
the discontinuity in flow pattern transition prediction. In addition, the unified modeling approach tends to be
realistic in its assumption that slug flow is the predominant flow pattern existing in a pipe, a pattern from which
other flow patterns develop. It has been found that this is an effective and successful approach for multiphase
flow modeling.
Thus, the current state of the art in multiphase flow in pipes is the emergence of both two-fluid transient
simulators and steady-state mechanistic models that more accurately describe the physical phenomena that occur.
Transient simulators have the capability of analyzing complex time-dependent problems but often suffer from
convergence problems. The improved technology also carries an additional cost. Both transient simulators and
mechanistic models are complex and require specialized training to understand and use. Interpretation of results is
better carried out by engineers with a specialized background who are fully aware of any simplifying assumptions
or limitations that have been included in the developments.
1.6.1 Petroleum Industry Future Challenges. As existing petroleum reservoirs mature, production of interstitial
and injected water imposes rheology, hydrodynamic, and flow assurance challenges, such as emulsions, complex
three-phase hydrodynamics, and corrosion. These challenges are further exacerbated when production of other
injected fluids used in chemical, thermal, and miscible floods is considered.
Furthermore, sufficient flow energy (pressure) to transport the required flow rates of hydrocarbons from reser-
voirs to processing facilities is a major issue in mature reservoirs. The use of artificial-lift methods such as electri-
cal submersible pumps and gas lift and the use of surface multiphase pumps may introduce flow assurance issues
such as creation of tight emulsions in water-producing wells or organic/inorganic solid formation resulting from
sudden heating and compression of fluids.
Major oil companies and large independent oil companies have sold most of their low-profit onshore fields to
smaller, independent oil companies. Majors are now concentrating on exploration and development of higher-
risk, more profitable reservoirs. Many of these discoveries are in deepwater regions in the Gulf of Mexico and
areas in the world such as Brazil and West Africa. Other target areas are heavy-oil onshore and offshore reser-
voirs with large potential reserves. All these target areas present serious technical challenges. For example, in
order to develop remote, deepwater offshore fields economically, a cost-effective approach often involves flowing
the wells to an existing platform through long multiphase flowlines (Oliemans 1994). In these flowlines, flow
assurance issues such as organic/inorganic solid formation and deposition (hydrates, paraffin, asphaltenes, scale),
high-viscosity cold flow, erosion/corrosion, terrain slugging, and flowline/platform/riser severe slugging become
important.
Onshore heavy-oil production and transportation also introduces several flow assurance issues relating to the
nature of the produced fluids and the recovery techniques used to extract oil from the reservoirs. The flow of
very-high-viscosity oils may impair the economic deliverability of wells as a consequence of complex and poorly
understood flow characteristics and inaccurate predictions using existing two-phase flow models (Gokcal et al.
2006). Emulsion formation with heavy oil complicates the rheology and downstream separation because of large
density and viscosity differences. In addition, heavy-oil production requires special recovery techniques that
affect the production and transportation of the oil. For example, the cold heavy-oil production with sand technique
allows the production of sand to maximize recovery. This, in turn, complicates the flow structure and can result in
eroding and/or plugging the system components. Cyclic-steam-stimulation and steam-assisted-gravity-drainage
recovery techniques result in water production and hot fluids that may cause tight emulsions, system corrosion,
and thermal fatigue. Therefore, heavy-oil system design and operation require appropriate modeling tools based
on sound physical understanding of the flow behavior to ensure environmentally safe and cost-effective design
and operation.
Increasing attention to environmental issues adds a major challenge to multiphase flow technology. For exam-
ple, zero gas flaring and carbon dioxide emission and protection of marine and land environment are required to
address today’s global concerns. Transportation of exhaust gases and unwanted produced fluids is a viable alter-
ative to protect the environment. Furthermore, to ensure system integrity and safety, a low tolerance of uncertainty
must be adopted, and this requires more-accurate and field-validated prediction models.
1.6.2 Multiphase Technology Future Direction. The future issues noted in the preceding subsection challenge
current multiphase technology. Great effort must be devoted to expand the current technology envelope in order
to cope with future challenges. Several technical issues are of high priority to the multiphase flow community. For
example, three-phase flow in pipelines and well tubing is extremely important to future multiphase phase technol-
ogy. Three-phase flow patterns, as well as water holdup and flow characteristics such as emulsions, dispersions,
and dispersed-phase droplet size distribution, are significant in the design of a transport system and downstream
processing facilities.
Mechanistic modeling of heat transfer that is dependent on flow pattern is critical in system thermal man-
agement to overcome flow assurance challenges of organic/inorganic solid formation/deposition. Formation of
asphaltenes, paraffins, hydrates, and scales must be predicted with a high degree of accuracy to design operation-
ally safe and cost-effective systems.
The main issue of heavy-oil multiphase technology is whether current conventional oil modeling is adequate.
Engineers must determine whether accuracies of existing pressure loss and liquid holdup predictions are accept-
able for economic design and reliable operation. A preliminary high-viscosity two-phase flow study by Gokcal et
al. (2006) showed current modeling capabilities to be inadequate. Therefore, existing two-phase flow mechanistic
modeling must be revised for use with heavy oil. For example, current flow pattern maps must be assessed accord-
ing to the hydrodynamics of heavy-oil, two-phase flows. Furthermore, new empirical closure relationships must
be developed to replace existing ones that were developed for conventional oils.
Modeling of both natural and imposed transients in multiphase flow is the least developed technology, yet it
is the highest in economic impact and environmental safety. Improved prediction of liquid slug behavior such as
growth, dissipation, and initiation along extremely long and large-diameter pipelines is required for the design
of slug catchers and separators. Prediction of liquid accumulation along pipelines is crucial for design of pigging
operations and corrosion inhibitor programs. Current transient models may not be capable of predicting severe
slugging for deepwater applications involving very long and large-diameter platform/riser systems. Improvement
of current transient slug-tracking models should be a future priority for safety reasons.
Fortunately, the petroleum industry and the multiphase flow technology community realize the upcoming future
challenges and are vigorously addressing them. Providing technically sound, environmentally safe, and economi-
cally profitable solutions to future challenges is never an easy matter. However, progress can be made and requires
collective efforts. For example, governmental and industrial funding of research is a key issue in this effort. More
importantly, efforts to find solutions require an interdisciplinary research approach. For instance, flow assurance
issues require input from physical chemists, chemical engineers, process engineers, and petroleum engineers
to find a comprehensive solution. Similarly, three-phase flow in pipes requires collaborative efforts of material
scientists, fluid dynamicists, and engineers. Although the future seems challenging and complex, it also looks
interesting and full of adventures. This book is a step toward solving these future challenges.
The basis for virtually all computations involving fluid flow in pipes is conservation of mass, momentum, and
energy. Application of these principles permits calculating changes in pressure and temperature along pipes. It
is important to understand the fundamentals of single-phase flow in pipes before discussing multiphase flow in
pipes. Therefore, this chapter is considered as an introduction to multiphase flow in pipes.
2.1 Objectives
The objectives of this chapter are to
• Present the fundamentals of single-phase flow in pipes for Newtonian and incompressible fluids
• Present the conservation laws of mass and linear momentum
• Develop a single-phase pressure gradient equation for Newtonian incompressible and compressible fluids
• Present non-Newtonian fluid flow behavior
2.2 Definitions
It is important to understand fluid physical behavior along pipes before predicting its flow parameters. The two
main fluid physical properties related to fluid flow in pipes are viscosity and density. On the basis of fluid viscos-
ity, fluids can be classified into two categories—namely, Newtonian and non-Newtonian. Similarly, fluids can be
classified into incompressible and compressible according to their density behavior.
2.2.1 Newtonian Fluid. A Newtonian fluid (named after Isaac Newton) is a fluid with constant viscosity, regard-
less of the shear force applied on it. This definition indicates that Newtonian fluid viscosity depends only on pres-
sure, temperature, and composition, and not on the shear forces. The constant viscosity behavior of a Newtonian
fluid is represented by the constant slope of a straight line through the origin of a plot of shear stress vs. shear rate.
Newton’s law for this type of (Newtonian) fluid can be written as
dv x
τ yx = − µ ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.1)
dy
where tyx = shear stress or momentum flux, Pa; dv x dy = shear rate, 1/second; and µ = viscosity, Pa·s. The nega-
tive sign in front of the viscosity indicates that shear stress (τ) transfers from high velocity to low velocity (i.e., in
the direction of velocity decrease). All gases and homogeneous nonpolymeric liquids follow Newtonian behavior.
Conversely, a non-Newtonian fluid is one that has an increasing viscosity trend (shear thickening) or decreasing
viscosity trend (shear thinning) with shear stress. Furthermore, a non-Newtonian fluid may have a constant vis-
cosity but require an initial shear force to flow (plastic flow). Fig. 2.1 shows Newtonian and non-Newtonian fluid
behaviors. A detailed description of non-Newtonian fluid flow is presented in a later section of this chapter. In oil
and gas production, crude oil, formation water, and natural gas are all considered to be Newtonian fluids. Other
fluids, however, such as oil/water emulsions or slurries (i.e., oil and produced solids) are often found to display
non-Newtonian behavior.
2.2.2 Incompressible Fluid. An incompressible fluid is characterized by a constant density with pressure
and temperature. Although there is no truly incompressible fluid, this assumption simplifies the mathematical
formulation and calculation significantly with minor error. Water is the best example of an incompressible fluid,
Plastic flow
Newtonian flow
τ (Pa)
0
dνx /dy (1/s)
whereas crude oil is considered a slightly compressible fluid. Conversely, natural gas is a compressible fluid
because of its low intermolecular forces that cause a density change with pressure and temperature.
2.2.3 Steady-State Condition. A steady state is a special flowing condition in which the mass, linear momen-
tum, and energy inflow rates in a pipe are equal to outflow rates; in other words, the rate change of mass, linear
momentum, and energy along a pipe is zero. In steady-state condition, the conservation laws are independent of
time. With the exception of Chapter 11, we consider that fluids in an oil and gas production system are Newtonian,
and flowing under steady-state condition.
2.3.2 The Continuity Equation. The law of conservation of mass (continuity equation) is shown in Fig. 2.2 and
can be physically written as follows.
∆
ρ Ap v x − ρ Ap v x + ∆x = ρ∆xAp. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .(2.3)
∆t
Dividing Eq. 2.3 by the pipe volume (Ap∆x), Eq. 2.3 becomes
ρ v x − ρ v x + ∆x ∆
= ρ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.4)
∆x ∆t
Reducing the pipe length to a differential element by taking the limit as Dx approaches zero, and using the defini-
tion of derivatives, Eq. 2.4 becomes
∆x
min CV mout
x x+∆x
d ( ρ vx ) dρ
− = .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.5)
dx dt
Eq. 2.5 is the 1D continuity equation that describes the time rate change of fluid density at any time and pipe
length. A special form of the continuity equation can be written for steady-state condition as follows.
d ( ρ vx )
= 0.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.6)
dx
For an incompressible fluid, Eq. 2.6 further reduces to the following.
dv x
= 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.7)
dx
Eq. 2.7 indicates that for a steady-state incompressible pipe flow, the fluid velocity is constant along the pipe.
2.4.2 Equation of Motion. The general conservation of linear momentum law (equation of motion) can be written
as follows.
For steady-state condition, Eq. 2.6 is applied, resulting in the rate of momentum accumulation in the control vol-
ume becoming zero. Thus Eq. 2.8 reduces to
The momentum rate terms in Eq. 2.9 can be mathematically defined as follows.
For incompressible flow in the control volume (i.e., Eq. 2.7 is applicable), the change of momentum rate across
the control volume is zero, thus reducing Eq. 2.9 to
Flow direction
τw
∆L p+∆p
τw
ρgsinθ
ρgcosθ
ρg
p
The gravity force is the weight of fluid and is the product of hydrostatic pressure ( ρ g∆L sin θ ) and pipe
cross-sectional area. The gravity force direction depends on the flow direction as follows:
a. In upward flow the gravity term is negative (i.e., pressure loss because of the positive inclination angle).
b. In downward flow the gravity force is positive (i.e., pressure gain because of the negative inclination
angle, which pulls the fluid in the direction of flow).
c. In horizontal flow, the gravity force is zero.
Simplifying and rearranging Eq. 2.11 to solve for the pressure-drop term gives
( )
∆pAp = ∆L −τ wπ d − ρ gAp sin θ .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.12)
The pressure change across the control volume (pressure gradient) can be solved from Eq. 2.12 as
∆p 4τ
= − w − ρ g sin θ .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.13)
∆L d
Eq. 2.13 determines the change of pressure across the pipe segment. To calculate the pressure change, evaluation
of the wall shear term (tw), which is associated with the kinetic behavior of the fluid, should be predicted first.
This can be accomplished by defining a dimensionless friction factor that is the ratio of the wall shear stress to the
kinetic energy of the fluid per unit volume (Bird et al. 2002), as follows.
τw
fF = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.14)
ρv2 2
where fF is the Fanning friction factor. Throughout this book, we adopt a Moody (1944) friction factor f, which
is four times larger than the Fanning friction factor. In terms of the Moody friction factor, Eq. 2.14 can be solved
for shear stress as follows.
f ρv2
τw = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.15)
8
Substituting Eq. 2.15 into Eq. 2.13, the final form of the pressure gradient equation is
∆p f ρv2
=− − ρ g sin θ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.16)
∆L 2d
2.4.3 Friction Factor Evaluation. Calculation of frictional pressure gradients requires determining values for
friction factors. Friction factor values depend on Reynolds number, pipe wall roughness, and flow regime in the
pipe (i.e., laminar or turbulent flow). Therefore, the procedure first requires evaluating whether the flow is laminar
or turbulent.
Reynolds Number and Flow Regime. The Reynolds number is a dimensionless number defined as the ratio of
inertial to viscous forces. The Reynolds number is defined as follows.
ρ vd
N Re = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.17)
µ
where ρ = kg/m3, v = m/s, d = m, and µ = kg/(m × s). Reynolds number is consistent for SI units. However, when
oilfield units are used (ρ = lbm/ft3, v = ft/sec, d = ft, and µ = cp), the units are inconsistent; thus, a unit conversion
coefficient of 1,488 is multiplied by the numerator of Eq. 2.17 to convert cp to lbm/ft-sec. If Reynolds number is
less than 2,000, flow is laminar. Otherwise, the flow regime is turbulent.
Pipe Wall Relative Roughness (ε /d). Relative roughness is the dimensionless roughness of the internal pipe wall
and is a measure of the texture of the surface. It is characterized by the vertical irregularities and deviations of a real
surface from its ideal form, as shown in Fig. 2.4. If these deviations are large, the surface is rough; if they are small,
the surface is smooth. The roughness is often affected by corrosive fluids and solid particles flowing in the pipe; thus,
it is a function of operation time, fluids, and pipe material. The relative roughness is the average actual roughness
divided by the pipe inside diameter. The roughness values of common material are illustrated in Fig. 2.5.
Friction Factor. Friction factor can be determined from the Moody (1944) chart shown in Fig. 2.5, which illus-
trates the flow regime regions and their respective friction factor behavior. Fig. 2.5 shows four flow regimes that
are a function of Reynolds number, pipe wall relative roughness, or both. These regime regions are as follows:
1. Laminar flow (0.1 < NRe < 2,000): characterized by low-velocity flow with the fluid streamlines being
straight lines. The Moody friction factor in the laminar region is a function of NRe only and can be deter-
mined analytically by equating the friction component in Eq. 2.16 to the pressure gradient defined by the
Hagen-Poiseuille equation for flow in a capillary tube (Bird et al. 2002).
64
f= . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.18)
N Re
2. Transition region (2,000 < NRe < 3,500): unstable region in which the laminar friction factor can be extrapolated.
0.1
0.09
0.08 Laminar Turbulent
Fully Turbulent
0.05
0.07
0.04
0.06 –2
2ε 0.03
f = 1.74 – 2.0log
0.05 d
Lam
0.02
f =N e
i
nar
0.015
0.04
64
Flow
0.01
R
0.008
d
ε
Relative Roughness,
0.006
Friction Factor,f
0.03
0.004
0.025 –2
2ε 18.7 0.002
f = 1.74 – 2.0log +
d NRe f
0.02 0.001
0.0008
0.0006
0.0004
0.015
Material Roughness (ft)
Smooth pipes 0.0002
Steel 0.003-0.03
Concrete 0.001-0.01 f = 0.184N –0.2 0.0001
Wood 0.0006-0.003 Re
Cast Iron 0.00085
Galvanized Iron 0.0005 0.000,005
Asphalted Cast Iron. 0.0004
0.01 Commercial steel 0.00015
Wrought Iron 0.00015
0.009 Drawn tubing 0.00006
0.008 0.000,001
3 4 5 6 8 108
103 2 3 4 5 6 8 104 2 3 4 5 6 8 105 2 3 4 5 6 8 106 1 2 3 4 5 6 8 107 2
ε
= 0.000,001 ε
d = 0.000,005
ρvd d
Reynolds Number, NRe =
µ
3. Turbulent flow (NRe > 3,500): characterized by high-velocity chaotic flow with random fluid particle
motion. The friction factor in the turbulent flow regime is classified into
a. Smooth-pipe friction factor (modified Blasius 1908)
−0.2
f = 0.184 N Re . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.19)
dp f ρv2
=− − ρ g sin θ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.22)
dL 2d
Eq. 2.22 is the celebrated pressure gradient equation for an incompressible steady-state flow condition. Eq. 2.22
comprises two components—namely, frictional pressure gradient and gravitational pressure gradient.
dp dp dp
= + . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.23)
dL T dL f dL g
The acceleration (kinetic) pressure term is neglected in Eqs. 2.22 and 2.23 because of the assumption of zero
change of momentum rate across the control volume (i.e., there is constant velocity and density) in the derivation
of the linear momentum conservation equation. The acceleration term is important in low-pressure systems (<100
psi), which are rarely encountered in petroleum production systems. Therefore, throughout this book, the assump-
tion of zero acceleration pressure term is adopted. For details on the acceleration term, refer to Shoham (2006).
2.5.1 Concept. The pressure gradient is the change of pressure per unit pipe length and is used to calculate the
pressure at any point along the pipe. The total pressure gradient can be negative, positive, or zero, depending on
the pipe inclination angle (θ) and the magnitude of the frictional and gravitational pressure-loss components. For
example, if the flow is horizontal or upward, the pressure gradient is negative, as in Eq. 2.22. However, if the flow
is downward (i.e., negative inclination angle) and the magnitude of the gravitational term is greater than or equal
to the frictional term, the pressure gradient is positive or zero, respectively.
2.5.2 Application. Fig. 2.6 shows a segment of a pipe with diameter d, length dL, and inclination angle θ. Assum-
ing isothermal conditions, and given flow rate, fluid properties, and the pressure at point 1, the pressure at point 2
can be determined by the pressure gradient as follows.
dp
p2 = p1 + dL = p1 + dp.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.24)
dL
For incompressible, isothermal, and steady-state flow conditions with a constant θ, the pressure gradient is con-
stant along the pipe because the density and velocity are constants. Therefore, evaluation of the pressure gradient
at point 1, where the pressure is known, can be applied all along the pipe with reasonable accuracy. Therefore,
dp dp dp
≈ ≈ ≈ constant.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.25)
dL 1 dL 2 dL
In this case, the pressure change along the entire pipe length is
dp
∆p2 −1 = L .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.26)
dL
Example 2.1—Single-Phase Liquid Pressure Gradient in a Pipeline. 4000 m3/d of water are flowing in a
9-km-long smooth pipeline. The pipeline inner diameter (ID) and inclination angle are 152 mm and +7°, respec-
tively. The pipeline outlet pressure is 2.07 MPa (300 psi). Determine the pipeline inlet pressure if the water viscos-
ity and density are 0.001 Pa·s and 1000 kg/m3, respectively.
Solution 2.1
dp kg m
= − ρ g sin θ = − 1,000 3 9.8 2 ( sin 7 )
o
dL g m s
kg
= −1.194 kPa m Pa=1 .
m ⋅ s2
2
L
dL
q θ
m3
4,000
q d
v= = = 2.55 m s.
s
Ap
(π 4 ) ( 0.152 m ) 86,400
2
= ( 0.184 )( 387,600 )
−0.2 −0.2
f = 0.184 N Re = 0.01403.
= − =− = −0.30 kPa m .
dL f 2d 2 ( 0.152 )
2. Total pressure gradient and inlet pipe pressure.
dp dp dp
= + = −1.194 − 0.30 = −1.49 kPa m.
dL T dL g dL f
dp
pinlet = poutlet − L p
dL T
= 2070 − ( −1.49 )( 9,000 )
= 15,480 kPa ( 2,245 psi ) .
15
L A− B = = 44 m.
sin ( 20 o )
15
LB − C = = 86 m.
sin (10 o )
∆p 860
= = 6.6 kPa m.
∆L 130
Since the change in pipeline elevation from the entrance to the exit of the pipeline is zero, the gravitational pres-
sure drop along the pipeline is zero. Thus, the total pressure drop reduces to the frictional pressure component.
dp dp f ρv2
= = − = −6.6 kPa m .
dL T dL f 2d
A C
q
15 m
θ1 θ2
B
2 d ( dp dL ) f
v= .
fρ
( 2 )( 0.1016 )( 6.6 )
v= = 0.275 m s.
( 0.02 )(883)
Ap = (π 4 ) ( 0.1016 ) = 0.0081 m 2 .
2
dp dp
≠ ≠ constant = f ( p, T ).. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.27)
dL 1 dL 2
Therefore, to determine the pressure gradient in an isothermal, compressible, single-phase flow, the pressure
gradient must be integrated with respect to pressure along the pipe. There are two common methods—namely,
Cullender and Smith (1956) and the average pressure method. In this book, the average pressure method is used
to numerically integrate the gradient equation. Cullender and Smith’s (1956) method was developed before com-
puter programs were available for these calculations and is no longer used in practice. For details on Cullender
and Smith (1956), refer to Brill and Mukherjee (1999).
2.6.1 Average Pressure Method. The average pressure method requires dividing the pipe into small increments,
as shown in Fig. 2.8. The average pressure, fluid properties, flow rates, and pressure gradient are assumed constant
in each increment, given that the length of each increment is relatively small.
In addition, the average pressure method requires a known temperature in each increment for the calculation of
fluid properties. The temperature in each increment can be predicted assuming that
1. The temperature distribution function along the pipe is isothermal (constant), or known
2. Or, a heat transfer model (energy equation) is coupled with the pressure gradient (momentum equation) to
predict the temperature gradient along the pipe.
In this part of the book, the first approach is adopted for its simplicity, leaving the heat-transfer model discussion
for Chapter 7. The average pressure method requires an iterative approach in each increment to determine the
pressure drop across each increment. Thus, the pressure drop is the sum of incremental pressure drops along a
pipe as follows.
n
dp
∆p2 −1 = ∑ dLi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.28)
i =1 dL i
i=n
L
3
1 2
i=1 dLi
q θ
Eq. 2.28 assumes a pipe segment with constant geometrical parameters (d and θ) and constant mass flow rate.
However, in the case of more complex pipe geometry, as shown in Fig. 2.9, the pipeline can be divided into pipe
sections with constant geometrical parameters and mass flow rate. Eq. 2.28 is applied for each pipe section, and
the pressure drop of each section is summed over the entire pipeline as follows.
m n
dp
∆pT = ∑ ∑ dLi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.29)
i =1 dL i
j =1
Calculation Procedure. Fig. 2.10 shows one pipe increment over which the calculation procedure for com-
pressible single-phase flow pressure gradient is carried out.
The calculation steps are listed as follows.
1. Read input data: geometrical parameters (d and θ), standard condition flow rate (qsc), fluid specific gravity
or composition, pipe inlet pressure (pi), and increment average temperature ( T ).
2. Guess pipe increment outlet pressure (pi + 1(G)) and calculate increment average pressure p = pi + pi +1(G ) 2. ( )
3. Calculate fluid physical properties, in-situ flow rate, and velocity at average pressure and average tem-
perature (Appendix B or C).
4. Determine friction factor (Fig. 2.5), average pressure gradient (Eq. 2.22), and increment outlet pressure
(Eq. 2.24).
5. Compare the guessed and calculated outlet pressures by calculating convergence criterion as
pi +1(G ) − pi +1(C ) pi +1(C ) ≤ ε , where ε is a small predefined value. Three cases will result:
a. Convergence is achieved: set pi + 1(C) as the inlet pressure (pi) of the next increment and march forward.
b. Convergence is not achieved: set calculated outlet pressure pi+ 1(C) as the new outlet pressure pi + 1(G) and repeat
Steps 2–5 until convergence is achieved. Convergence should be achieved in less than 10 iterations.
c. Number of iterations exceeds 10, and no convergence is achieved: numerical problem occurs; stop
iteration and proceed with unconverged pressure gradient value in this increment.
6. When convergence is achieved, march forward to the next increment. Take the outlet pressure of the previ-
ous increment as the inlet pressure of the new increment and repeat Steps 2–5. Repeat this procedure for
all pipe increments in a pipe section.
7. Repeat Steps 1–6 for each pipe section in the piping system to determine the pressure distribution along
the entire system.
q2
q1 j=4
i=1 2 n
j=3
j=1 j=2
dL
p
pi pi+1
T
Fig. 2.11 is a flow chart summarizing the compressible flow calculation procedure for one pipe increment.
Start
Guess pi+1(G)
Calculate
increment p
Calculate fluid
properties, in-situ
flow rate and
velocity at
p and T
Convergence No
pi+1(G) = pi+1(C)
on pi+1
Yes
Next Incrementt
Fig. 2.11—Marching computational algorithm flow chart for one pipe increment.
Example 2.3—Compressible Single-Phase Flow in a Vertical Pipe. Calculate the flowing bottomhole pressure
in a vertical gas well using the average pressure method with two equal pipe increments. The following data are
known.
Preliminary Calculation
T = 62 °C
Ap = (π 4 ) d 2 = (π 4 ) ( 0.062 ) = 0.003 m 2 .
2
T1 = 62 °C (143.7°F).
T2 = 99.5°C (211°F).
T = 99.5 °C
Tbh = 118.3 °C
13.83 + 15.56
p= = 14.7 MPa (2,125 psia)
2
2. Calculate fluid physical properties, in-situ flow rate, and velocity at p and T .
Z = 0.766.
pM w
ρ= = 149 kg/m 3.
ZRT
µ = 0.000018 Pa·s.
ZT ( 0.766 )( 62 + 273)
Bg = 352.3 = 352.3 = 0.00615 m3/std m3.
p (14.7 × 10 6
)
q = qscBg = 861 m3/d.
861
v= = 3.32 m s .
( 0.003)(86,400 )
3. Calculate friction factor, average pressure gradient, and increment outlet pressure.
dp f ρv2
=− − ρ g sin θ = −192 − 1460 = −1,652.2 Pa m.
dL 2d
dp
∆p = L = ( −1,652.2 )(1,524 ) = −2.5 MPa.
dL
p2 = 13.83 + 2.5 = 16.3 MPa (Not close enough to initial guess).
4. Set p2(C) = p2(G) = 16.3 MPa and repeat Steps 1–3 as a second iteration.
p = 15.06 MPa (2,176 psia).
Z = 0.766.
pM w
ρ= = 153 kg/m 3 .
ZRT
µ = 0.000018 Pa·s.
ZT ( 0.766 )( 335)
Bg = 352.3 = ( 352.3) = 0.006 m3/std m3.
p (15.06 × 106 )
q = qscBg = 840 m3/d.
v = 3.24 m/s.
ρ vd (153)( 3.25)( 0.062 )
N Re = = = 1.71 × 10 6.
µ 0.000018
dp f ρv2
=− − ρ g sin θ = −193 − 1,499 = −1,692 Pa m.
dL 2d
dp
∆p = L = ( −1,692 )(1,524 ) = −2.58 MPa.
dL
1. Increment inlet pressure is 16.41 MPa. Using the initial guessed pressure gradient of 1.134 kPa, the
guessed increment outlet pressure is 18.14 MPa, which gives
2.7.1 Non-Newtonian Fluid Characterization. A general correlation to predict the viscosity of a non-Newtonian
fluid is difficult to obtain because of the unique characteristics of each non-Newtonian fluid. Therefore, to accu-
rately determine the viscosity behavior of a non-Newtonian fluid, a laboratory experimental measurement must
be conducted on the fluid using an appropriate viscometer. Dispersed oil/water flow in pipes can be treated as a
non-Newtonian fluid with the following assumptions:
1. The mixture is homogeneous (slippage is neglected). Thus, in-situ liquid holdups are the same as their
respective input volume fractions (no-slip liquid holdup).
2. The rheological behavior of the oil/water mixture is suitably described by the Ostwald–de Waele power
law model (Bird et al. 2002) as follows.
If n is unity, Eq. 2.30 describes Newtonian behavior, and K will be equal to the constant viscosity, µ. For a typi-
cal oil/water mixture, n will usually be less than unity, and Eq. 2.30 will describe pseudoplastic (shear-thinning)
behavior. It is also possible for an oil/water mixture to have n greater than unity, resulting in dilatant (shear-thick-
ening) behavior. The type of behavior that a fluid system will follow is normally unknown and can be determined
from laboratory experiments using an appropriate viscometer. To determine n and K, a logarithmic plot of wall
shear stress τ w = ( dp dL ) d 4 vs. nominal shear rate (γ w = 8 v d ) is required, where the slope is n′ and intercept
is K ′. The nominal shear is the true shear rate at the pipe wall or the shear rate for a Newtonian fluid. It is used here
instead of the true wall shear rate for scaling up from small- to larger-diameter pipelines, given that the laminar
flow regime exists and no “wall-slip” occurs, which is rarely significant for d > 50 mm. It is important to note that
the n′ and K ′ determined from the characteristic curve correspond to wall shear stress and shear rate, not to the
true shear stress and shear rate. It has been shown from analogy between the power law and non-Newtonian true
shear rate/shear stress relationship that n = n ′and K = K ′ [( 3n + 1) 4 n ] .
n
Non-Newtonian Reynolds Number. Metzner and Reed (1955) introduced the concept of a generalized Reyn-
olds number for non-Newtonian flow. The final expression of their generalized Reynolds number in terms of n
and K is as follows.
ρ v 2− n d n 4n
n
′ =
N Re . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.31)
8n −1 K 1 + 3n
If n is unity, Eq. 2.31 reduces to the usual Reynolds number definition with K as µ.
Laminar-Turbulent Flow Transition. A number of methods have been presented in the literature to predict the
critical value of Reynolds number at which transition to turbulent flow occurs. For a power law fluid, Ryan and
Johnson (1959) proposed the following expression for the transition from laminar to turbulent flow in terms of
flow behavior index n.
If the Reynolds number calculated from Eq. 2.31 is less than the critical Reynolds number from Eq. 2.32, the flow
regime is laminar; otherwise, the flow is turbulent.
Non-Newtonian Friction Factor. In general, the procedure for calculating friction factors for non-Newtonian
fluids is similar to that for a Newtonian fluid. Several friction factor equations are presented here with respect to
flow regimes and pipe wall roughness.
1. Laminar flow. Using the non-Newtonian Metzner and Reed (1955) generalized Reynolds number, the
laminar non-Newtonian Moody friction factor is as follows.
64
f′= .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.33)
′
N Re
2. Turbulent flow through smooth pipes. Dodge and Metzner (1959) proposed the following implicit Moody
friction factor equation.
{ }
−2
4.0 0.4
fs′ = 4.0 0.75 log N Re
′ fs′ ( ) − 1.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.34)
1− n 2
n n
3. Turbulent flow through rough pipes. Govier and Aziz (1972) suggested the following Moody friction fac-
tor for power law pseudoplastic fluids flowing in rough pipes.
fr
f ′ = fs′ Newtonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.35)
fsNewtonian
where fs′ = friction factor calculated from Eq. 2.34, frNewtonian = friction factor for rough pipe calculated from
Eq. 2.20 at the same generalized Reynolds number, and fsNewtonian = friction factor for smooth pipe calculated from
Eq. 2.19 at the same generalized Reynolds number.
4. Fully turbulent flow through rough pipes. Govier and Aziz (1972) recommended the following expression.
−2
d 2.65
f ′ = 4 4.06 × log + 6.0 − . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.36)
2ε n
2.7.2 Non-Newtonian Frictional Pressure Gradient. Once the friction factor is determined for the non-Newto-
nian fluid system, the pressure gradient can be calculated from Eq. 2.22, which is slightly modified in the friction
factor term as follows.
dp f ′ρv2
=− − ρ g sin θ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2.37)
dL 2d
Example 2.4—Non-Newtonian Oil/Water Mixture Pressure Drop. A horizontal pipe discharges a dispersed
oil/water mixture at atmospheric pressure. Determine the pressure at the inlet of the pipe required to maintain
a constant volumetric flow rate of the dispersion for input water fraction fw = 0.2. The following are also given:
ρ v 2− n d n 4n
n
′ =
N Re
8n −1 K 1 + 3n
890 × 4.22 − 0.8589 × 0.5080.8589 4 × 0.8589
0.8589
= .
80.8589 −1 × 0.33156 1 + 3 × 0.8589
= 9,995.
3. Calculate the laminar-turbulent boundary.
=
(1 + 3 × 0.8589 )2
= 3,312.
′ > 3,312, the flow regime is turbulent.
Since N Re
4. Compute the friction factor for turbulent flow of pseudoplastic fluids in rough pipes using Eq. 2.35. The
turbulent smooth-pipe friction factor is calculated iteratively from Eq. 2.34.
−2
fs′ = 4.0
4.0
( 0.8589 )
{
0.75 Log 9,995 × f s′
1− ( 0.8589 2 )
− }
0.4
( 0.8589 )1.2
.
Using a trial-and-error procedure to solve the implicit friction factor, a final answer is found as
fs′ = 0.02278 .
To determine the rough-pipe friction factor from Eq. 2.35, the Newtonian smooth- and rough-pipe friction factors
from Eqs. 2.19 and 2.20, respectively, are calculated using the generalized Reynolds number.
fsNewtonian = 0.03073.
frNewtonian = 0.03477.
fr
f ′ = fs′ Newtonian = 0.02278 (1.1315) = 0.02577.
fsNewtonian
dp
p1 = p2 − dL
dL
= 101.325 − ( −0.398 )( 3,050 ) (191.4 psia).
= 1,315.22 kPa.
2.8 Chapter Key Learnings
• Single-phase incompressible and steady-state fluid flow in a pipe has a constant fluid velocity along the pipe.
• Conservation of mass and linear momentum laws can be coupled to derive the incompressible, steady-state,
single-phase pressure gradient equation in pipes.
• An incompressible, steady-state, single-phase pressure gradient equation is used for compressible flow in
pipes using the average pressure method to account for variation in fluid properties along the pipe.
• A non-Newtonian fluid has a nonlinear viscosity trend, which requires a special characterization to deter-
mine flow behavior parameters. A generalized Reynolds number and friction factor are determined for the
non-Newtonian fluids to calculate frictional pressure gradient.
2.9 Problems
Problem 2.1—Flow Regime Determination. What is the flow regime (i.e., laminar or turbulent) for 0.0019 m3/s
of fresh water flowing in a 76-mm-ID pipeline?
Problem 2.2—Reynolds Number. Derive an expression for Reynolds number in terms of q, d, and ν (kinematic
viscosity).
Problem 2.3—Crude Oil Pressure Gradient. A crude oil flows at a velocity of 1.524 m/s through a 50.8-mm-ID
pipeline with a head loss of 5.5 m in 30.5 m. What will be the head loss when the velocity is increased to 3.05 m/s,
given that the oil specific gravity and kinematic viscosity are 0.92 and 9.3 × 10–5 m2/s, respectively?
Problem 2.4—Water Pressure Gradient. The mean velocity of water flowing in a 300-mm-ID and 0.002-rela-
tive-roughness pipeline is 3 m/s. Given that the kinematic viscosity of water is 9 × 10–7 m2/s, determine the pres-
sure drop in kPa along a 300-m-long pipe.
Problem 2.5—Pipeline Sizing. A smooth 61-m-long pipeline is to transport water at 0.003 m3/s between two
tanks open to the atmosphere and with a surface elevation difference of 1.5 m. What is the required pipe diameter?
Problem 2.6—Pipeline Capacity. Calculate the flow rate for fresh water flowing through a 6-m-long, 50-mm-ID
horizontal smooth pipe with a head loss of 0.3 m.
Problem 2.7—Liquid Viscosity Determination. A liquid (go = 0.85) flows in a 101.6-mm-ID commercial steel
pipe at a rate of 0.0043 m3/s. If the pressure gradient over a 61-m-long horizontal pipe is 1.73 kPa, determine the
liquid viscosity.
Problem 2.8—Relative Roughness Estimation. Fresh water flows in a smooth pipeline at a Reynolds number of
106. After several years of operation, it is observed that one-half the original flow rate produces the same pressure
gradient as the original flow rate. Estimate the relative roughness of the deteriorated pipeline.
Problem 2.9—Pipeline Leaking. Water flows in a 0.3-m-ID pipeline with a constant friction factor of 0.025. The
pipeline is found to leak water at a point where the upstream and downstream pressure gradients of the leaking
location are 7 and 6.5 kPa/m, respectively. Determine the water-leakage rate.
Problem 2.11—Flow in Wells. Well 1 and Well 2 are identical wells and are producing from the same reservoir.
Well 1 produces through the tubing, whereas Well 2 produces through the casing/tubing annulus, as shown in
Fig. 2.12. Given that the cross-sectional area of the production tubing in Well 1 and the annulus in Well 2 are equal
(AT = Aan), determine which well has the higher flow rate? Support your answer.
q (m3/min) Dp (kPa)
0.047 16.389
0.249 25.793
0.711 35.059
1.304 40.314
2.394 47.990
4.100 55.735
6.020 61.129
7.382 67.352
Table 2.1—Viscometer data (Problem 2.10).
Tubing Tubing
Casing Casing
Well 1 Well 2
Problem 2.12—Liquefied Petroleum Gas (LPG) Storage in Salt Cavity. Salt cavities are often used for storage
of liquid petroleum gas (LPG) (Fig. 2.13). There are two possible salt cavity operating configurations: compen-
sated (with brine) and uncompensated (without brine). Weep holes are used to detect the presence of LPG in the
brine for compensated storage cavities. Anytime LPG is being injected into the cavity under normal conditions,
brine is produced up the tubing, with some of the brine entering the tubing at the tubing shoe and the remainder
entering the tubing through the weep hole. The amount of brine entering the tubing through the weep hole is small
and depends on the size of the hole and the pressure difference across the weep hole. Estimate the pressure differ-
ence across the weep hole for the following case, assuming the LPG level is above the weep hole.
LPG injection rate = 3180 m3/d ρB = 1027 kg/m3
dt = 0.203 m f = 0.02
Tubing shoe to weep hole length = 30.5 m.
Assume the LPG and brine velocities in the salt cavity are very small (negligible).
Brine
LPG
Tubing
LPG LPG
Weep hole
Brine
Tubing shoe
Brine
Problem 2.13—Giraffe Blood Circulatory System. The circulatory blood system of the giraffe is unique and
scientifically interesting. A giraffe’s heart can weigh more than 11 kg, is 0.61-m long, and must generate approxi-
mately double the blood pressure required for a human to maintain an average blood flow rate to the brain at 10
L/m. It is also known that a giraffe’s blood density and viscosity average approximately 1060 kg/m3 and 0.0018
Pa·s, respectively, and its heart and brain are connected by a smooth 3-cm-ID and 2.74-m-long blood vessel
(because of the long neck). Amazingly, the giraffe’s upper neck has several valves to prevent excessive blood flow
and pressure in the brain when the giraffe bends its neck below its heart level. Using single-phase fluid mechanics
principles, answer the following questions:
1. If the required average blood pressure in a giraffe’s brain is 120 mm-Hg, determine the required average
blood pressure that the giraffe’s heart must supply as the giraffe extends it neck vertically to eat from a
tree. Assume the giraffe’s heart is located just below its neck.
2. Determine the required pressure drop across a giraffe’s upper neck valve to maintain the blood pressure in
its brain at 120 mm-Hg when the giraffe bends its 2.74-m-long neck to drink from a water pond. Assume
the giraffe bends its neck in a ∩-shape with 0.61-m vertical upward and 2.13-m vertical downward neck
segments to reach the water pond.
Problem 2.14—Mass Balance in Expanded Pipe. A single-phase compressible mixture is flowing in a horizontal
expanding pipe from a small diameter to a large diameter, as shown in Fig. 2.14. Given the following information,
and neglecting the accelerational pressure gradient component at Point 2, determine the pipe diameter at Point 1.
Point 1 Point 2
vg = 5.5 m/s dp/dL = 113.3 Pa/m
rg = 220 kg/m 3
rg = 125 kg/m3
d1 = ? f = 0.02
1 2
d2 = 76 mm