Chemical Engineering Science: Abdalellah O. Mohmmed, Hussain H. Al-Kayiem, A.B. Osman

Chemical Engineering Science 238 (2021) 116611
Contents lists available at ScienceDirect
Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces
Investigations on the slug two-phase flow in horizontal pipes: Past,

presents, and future directives
Abdalellah O. Mohmmed a, Hussain H. Al-Kayiem b, A.B. Osman b
a
College of Engineering, Emirates Aviation University, 53044 Dubai International Academic City, Dubai, United Arab Emirates
b
Mechanical Engineering Department, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak, Malaysia
g r a p h i c a l a b s t r a c t
a r t i c l e i n f o a b s t r a c t
Article history: Precise prediction of the gas–liquid slug flow is crucial for proper design and operation at various industrial
Received 29 October 2020 processes. This paper provides a comprehensive review of published experimental and numerical works in
Received in revised form 24 February 2021 horizontal pipes, including a critical analysis of state-of-the-art parameters used to characterize the slug
Accepted 22 March 2021
flow regime. The discussed slug characteristics are slug length, slug frequency, and slug velocity.
Available online 26 March 2021
Throughout the paper, an endeavor does devote to discerning the areas where modern development work
is needed. Accordingly, some observed gaps in the literature, comprising hydrodynamic behavior at the
Keywords:
transition zone and the interaction between slug flow and structural pipe. The lack of comprehensive lab-
Two-phase slug flow
Behavioral slug characteristics
oratory experimental information for high liquid viscosity and the viscosity impact on the slug character-
Experimental istic is also observed. On the contrary, robust and less empirically independent CFD models are required for
CFD accurately predicting the slug characteristics with a minimal computational cost.
Horizontal pipe Ó 2021 Elsevier Ltd. All rights reserved.
1. Introduction
Among the potential two-phase flow regimes in a horizontal

pipe, intermittent flow occurs over a broad spectrum of gas–liquid
E-mail addresses: [email protected] (A.O. Mohmmed), hussain_kayiem@utp.
superficial velocities. The intermittent flow in horizontal pipes
edu.my (H.H. Al-Kayiem), [email protected] (A.B. Osman)
https://doi.org/10.1016/j.ces.2021.116611
0009-2509/Ó 2021 Elsevier Ltd. All rights reserved.
A.O. Mohmmed, H.H. Al-Kayiem and A.B. Osman Chemical Engineering Science 238 (2021) 116611
appears as the bulk of liquid regions fill the entire pipe (plugs or simplified by assuming the slug flow as a sequence of identical
slugs) followed by elongated gas bubbles or stratified flow regions. cells known as the unit cell concept. The unit cell concept takes
Accordingly, the intermittent flow is classified into slug and elon- steady fluid velocity and fully developed flow. Besides, it requires
gated bubble flow regimes. When the flow is stable and the liquid four closure laws to complete the solution process, and the solu-
slug is non-aerated, the flow pattern is termed as elongated bubble tion accuracy is partially relying on these closures. However, these
flow. As the flow rates increased, the slug liquid entrained with gas closure laws are abundant with unequal values, as reported by
bubbles, and the flow is termed as slug flow. Despite the distinc- (Taitel and Barnea, 1990, Fabre and Liné, 1992, Dukler and Fabre,
tion between these two flow regimes, the general intermittent flow 1994). These closures were developed from various flow condi-
is often known as slug flow (Taitel and Barnea, 1990). Slug flow is tions, pipe diameter, and fluid properties, consequently generate
an extremely complex and unsteady flow regime, even when the variant results under identical slug flow conditions. The initially
superficial gas and liquid velocity are kept constant at the pipe introduced unit cell concept (Dukler and Hubbard, 1975) has been
inlet. During stable slug flow, liquid shed at the slug tail is similar used widely for modeling the gas–liquid slug flow. The slug flow is
to the liquid picked up at the slug front. Thus, the developed slug divided into two structures: the slug liquid, which carries the liq-
length remains constant as the slug travel along the pipe (Fabre, uid phase, and an elongated bubble that flows over the liquid film.
2003). The wall shear stresses deaccelerate the liquid shed at the This concept assumes these two structures appear in the pipe suc-
back and forms a stratified layer for horizontal or near-horizontal cessively and repeatedly, as shown in Fig. 1. Although several
pipes. By contrast, for vertical or near-vertical pipes, the liquid authors, for instance (Dukler and Hubbard, 1975, Mishima and
picked up to accelerate and developed a falling annular film as Ishii, 1980, Ruder et al., 1989, Zhang et al., 2000, Soleimani and
the slug moves downward. The elongated bubbles separated the Hanratty, 2003, Taitel and Dukler, 1976, Ujang et al., 2006,
slug liquid have a cylindrical interface shape in the vertical flow Sanchis et al., 2011, Nieckele et al., 2013) have investigated the
and a nearly flat interface in the horizontal flow (Fabre, 2003). mechanisms of slug formation, limited studies focused on slug flow
Slug two-phase flow is a complex phenomenon occurred in var- characteristics as reported by (Dinaryanto et al., 2017). The con-
ious industrial applications and processes, such as oil and gas, temporary trend moves toward the formulation of approximate
nuclear power plants, and chemical industries. The existence of slug models capable of simulating the slug characteristics such as the
flow causes severe problems for designers and operators. The slug pressure drop, slug transitional velocity, and slug length accurately
liquids encompass high momentum that creates significant forces and confidently. A sophisticated slug flow model computes the
at the elbows, tees, or other process equipment. These force fluctu- slug liquid length, bubble length, slug frequency, slug initiation,
ations induce vibration, which consequently causes cyclic fatigue slug transitional velocity, void fraction, wall shear stresses, and
damage (Wang et al., 2018) or even pipe rupture, as reported by pressure drop (Fabre, 2003). Among these parameters, the slug ini-
(Gong, 2010). Furthermore, the low frequencies of slug flow can res- tiation, void fraction, slug length, slug frequency, and pressure
onate with the fundamental natural frequency of large piping struc- drop are crucial characteristics that provide clear insight into the
tures, which cause severe damage due to vibrational fatigue. slug flow behavior.
Besides, the elongated slug liquid contains a large quantity of liquid, This study strives to provide a thorough review of the slug flow
which results in flooding at the separators and damage to the other analysis and investigations in horizontal pipes involving the mod-
process facilities (Fabre, 2003). Physically, the slug flow regime con- eling, prediction of slug characteristics, and the impact of various
solidates the properties of the single-phase fluid flow (shear forces, parameters on the slug flow behavior. In what follows, this study
inertia forces, and pressure forces), along with the two-phase flow aims to demonstrate the previous arts, advancement, and develop-
(interface forces, contact angle, mass transfer, heat, and momentum, ment of the slug flow analysis chronologically.
etc.) (López et al., 2016). Due to these complexities and the encoun-
tered problems associated with slug flow, extensive studies have
been conducted to investigate this phenomenon in various applica- 2. Slug flow formation and transition
tions and structures.
The prediction of fluid properties such as pressure drop, inter- In the two-phase flow, the transition from one regime to another
face forces, heat and mass transfer, and momentum is a tedious depends on fluids properties, pipe inclination angles, superficial
task due to the unsteady nature of slug flow. There are significant velocities, etc. Among all, stratified to slug flow transition is crucial,
challenges encountered while modeling the hydrodynamic slug as the slug induces a transient internal pressure, which causes sev-
flow, namely: the dynamic behavior of the interface, which sepa- ere damage to the pipe structures. At low liquid velocities, the liquid
rates the gas–liquid in the pipes, gas entrainment on the slug liq- occupies the bottom of the pipe. The gas occupies the upper section
uid, and liquid droplets into the gas phase, besides the existence due to the gravitational effect; the resulted flow regime is called a
of the turbulence effect near the vicinity of the interface layer. stratified or a wavy–stratified regime. As the liquid velocity increase,
Owing to these challenges, precise prediction for the slug charac- the interfacial waves propagate and form a slug liquid. The formed
teristics is beyond the capabilities of the existing 1D models (Mo liquid slug fills the entire pipe cross-section while the gas occupies
et al., 2014). Although there are several attempts to analytically small voids inside, and the slug liquid is depicted as an aerated slug.
scrutinize the impact of the slug characteristics on the solid struc- The aeration level is highly dependent on gas velocities (Vaze and
tures (Li et al., 2020, Wang et al., 2018), obtaining an exact solution Banerjee, 2011). Considerable theoretical work has been carried
of the continuum govern equations has not achieved yet. There- out in the past few decades to understand the mechanism of the
fore, a diversity of approximate techniques have been adopted to transition to intermittent flow. One approach is to examine the sta-
assess the hydrodynamic slug characteristics. The former methods bility of the flow. Ancient endeavors by (Kordyban and Ranov, 1970)
utilized the empirical correlations derived from the ancient and (Wallis and Dodson, 1973) discovered a stability mechanism
researchers’ experimental results to the current advanced compu- whereby the slugs developed due to wave growth due to infinitesi-
tational methods such as image processing and Computational mal disturbances at the interface between the gas–liquid phases.
Fluid Dynamics (CFD) techniques. Hence, in the recent survey, Years later, (Taitel and Dukler, 1976) proposed an approach widely
the evolutionary voyage of predicting and determining the slug utilized to predict the transition to the slug flow regime. In their
characteristics is thoroughly discussed. approach, the shear stresses were neglected, and the non-linear
Modeling the slug flow entails a keen understanding of various effect of wave pressure fluctuations at the interface was considered
fundamental principles. However, the modeling complexity is using inviscid Kelvin–Helmholtz (IKH) instability. (Taitel and
2
(a)
Density Density Density
Slug Pocket Slug Liquid
Slug Liquid Slug Pocket Slug Liquid Slug

Pocket
Flow Direcon Idealized Slug Unit
(b)
Fig. 1. Slug flow in a horizontal pipe. (a) Idealized slug flow (Mohmmed et al., 2019); (b) Actual typical slug two-phase flow (Al-Kayiem, 2017 #18).
Dukler, 1976) concluded that the transition from stratified to inter- information regarding slug and roll wave flow regimes’ behavior in
mittent flow could be identified through one or more of three crite- terms of the relation between the slug formation and superficial
ria: stability of slug flow, viscous linear instability for long- gas–liquid velocities. (Kadri et al., 2009a) concluded that the slug
wavelength disturbance, and Kelvin–Helmholtz (KH) instability formation time decreased with increasing the superficial liquid
for stratified flow. Kelvin–Helmholtz for Viscous fluid (VKH) has velocity. While, at low superficial gas velocity, slug formation time
been used by (Wallis, 1969, Lin and Hanratty, 1986, Wu et al., increase with increasing the gas velocity. Besides, increasing the
1987, Andritsos et al., 1989, Barnea, 1991, Crowley et al., 1992) to pipe diameter leads to larger axial growth, consequently minimiz-
determine the stability of the stratified two-phase flow considering ing the probability of long slug appearance. In continuation of his
the wall shear stress with the two fluids model. As reported by previous study, (Kadri et al., 2009b) explained that an elongated
(Barnea and Taitel, 1993a), both IKH and VKH provided similar slugs liquid was formed from the development of gravitational
results for high viscous flow while a large discrepancy between waves at low superficial velocities. They recommended consider-
IKH and VKH for the low viscous fluids was observed. Also, at the ing the interfacial friction factor, gas entrainment, and the wave-
low amplification zone, (Barnea and Taitel, 1993a) confirmed that length to foster the model accuracy.
the slug flow or roll wave region was located between the neutral The stability of the stratified two-phase flow in horizontal pipes
stability line of the VKH and IKH analysis. For the high amplification was examined analytically by (Barmak et al., 2016), considering
zone, either slug or annular flow regimes were situated above the the potential wave numbers. Comparing their experimental
IKH neutral stability line. Moreover, two mechanisms for the transi- results, they concluded that at relatively low flow rates, the gas–
tion from stratified to slug flow were proposed (Woods and liquid stratified flow stability was observed controllable. Besides,
Hanratty, 1999). (1) at subcritical flow rate where the superficial using previous long-wave analysis, (Barmak et al., 2016) could
gas and liquid velocities are shallow, the slug liquid formed due to not predict their results, which further affirms the discrepancies
large amplitude gravitational waves; whereas. (2) at supercritical of the developed analytical models are high yet. In contrast, a mas-
flow rates, the slug liquid formed due to the roll-waves coalescence. sive difference between the published results was observed. Fur-
Furthermore, the theory of viscous long wavelength (VLW) was uti- thermore, extensive experimental investigations were conducted
lized to predict the slug flow transitions at low gas velocities. This by (Kong et al., 2018) in small and large pipe diameters to investi-
theory is capable to track and trigger the growth of long wavelength gate the air–water slug flow’s interfacial behavior utilizing four-
waves and their transition into a slug as reported by (Woods et al., sensor conductive probes. Their results indicated that the pipe
2000). Later, (Hurlburt and Hanratty, 2002) recommended that the diameter has immense effects on the interfacial structure of the
transition to the slug flow region can be predicted accurately using slug flow. At the same time, the bubble size increased with increas-
VLW model for low superficial gas velocities, while the for high ing the pipe diameter. On the other hand, the constant electric cur-
gas velocities, slug stability model performed better. Besides, they rent method (CECM) was utilized by (Rahmandhika et al., 2018) to
suggested to include the interfacial friction factors, hydraulic gradi- investigate stratified to slug flow transition in a horizontal pipe.
ents, bubble velocity, the void fraction, and the slip ratio over a wide Based on their work, different stratified to slug flow transition
spectrum of fluid properties for better prediction. mechanisms existed for various superficial gas (JG) and liquid
Correlations of the interfacial friction factors in air–water two- velocities (JL). They found that the onset of slugging was formed
phase flow have been developed in the previous works, e.g., by the wave coalescence mechanism at high JG. While at low JG
(Andritsos and Hanratty, 1987b, Bontozoglou and Hanratty, 1989, and JL, the slug formation was observed fast, and the transition
Simmons and Hanratty, 2001). wave transition model was devel- was similar to the previous outcomes of (Lin and Hanratty, 1986,
oped by (Kadri et al., 2009a) to predict the transition from strati- Andritsos and Hanratty, 1987b). They concluded that there is an
fied to slug or roll wave flow regimes for various superficial gas– agreement between most researchers about the transition mecha-
liquid velocities and pipe diameters. This model provided intrinsic nisms at low superficial velocities. However, at higher superficial
3
gas velocity JG > 3 m/s, the transition to slug flow is different coalesce into one due to gas carry-under between the slugs, which
among the previous studies. Following the previous research, entrained the entire gas between the slugs into the precursor slug.
(Rahmandhika et al., 2019) extended their work on the transition Besides, they reported that the mean slug length continues to
from stratified to slug flow regime. They concluded that the slug increase as the slug travel inside the pipe due to slug death and
formed due to wave growth mechanism at JG 1.88 m/s and slug coalescence mechanisms. Moreover, the slug length varies
0.05 m/s JL 0.2 m/s. Also, for the JG 2.83 m/s, the slug formed within wide spectrums around a mean value, unlike the other
due to wave coalescence. Their experimental results were under- characteristics of slug flow. Thus, it has been predicted using
ated using the theory of slug stability, although the comparison empirical correlations and analytical methods. The slug flow and
provided similar trend. While for the comparison with K H instabil- the film mixing process in vertical and horizontal pipes were inves-
ity theory, the theory overrated the experimental data, and the tigated by (Barnea and Brauner, 1985) utilizing a wall jet. They dis-
agreement was achieved for the cases where JG > 6.2 m/s. Even covered that the tail behind a short slug moved with higher
though there were massive data obtained from the experimental velocity than a tail behind a fully developed velocity profile. Conse-
and analytical investigation for the interfacial structure of slug quently, this causes the tails of short slugs to overtake their fronts
flow, the mechanism of slug initiation and formation is not fully and dissipate into a form of waves while moving inside the pipe. In
understood. Therefore, ancillary experimental investigations horizontal flow, (Barnea and Brauner, 1985) found that the mini-
regarding the hydrodynamic behavior at the transition zone are mum stable length that can propagate on the pipe without dissipa-
required. Additional experimental data is necessary to enhance tion was 32D. While (Dukler et al., 1985) found that the minimum
the interfacial friction factor’s accuracy ratio to the stratified inter- stable slug length is approximately 20D, it is in reasonable agree-
facial factor (Rahmandhika et al., 2019). Based on these experimen- ment with the finding (Barnea and Brauner, 1985). The utilized
tal data, supported correlations for the previously developed technique assumed the short slugs’ tails move faster than in the
theories were developed, consequently resolving the associated long slugs because short slugs’ velocity profile is not fully devel-
ambiguity with slug initiation, formation, and transition. oped. However, this outcome contradicts the observations
(Fagundes Netto et al., 1998), who concluded that slug length var-
ied with tail velocity randomly.
3. Slug flow characterization
The mean slug liquid length was found within the range of 12D-
30D by (Dukler and Hubbard, 1975, Nicholson et al., 1978) from their
As indicated by (Sharma et al., 1998), the slug length, slug fre-
experimental investigations in horizontal pipes. For small diameter
quency, volume fraction distribution, slug velocity, momentum,
pipes, (Barnea and Taitel, 1993b) concluded that the mean and max-
and energy transfer at the interface are the most crucial parame-
imum slug lengths were 1.5 and 3 times the minimum stable slug
ters for characterizing the slug flow. These slug characteristics
length, respectively. Whereas (Dukler et al., 1985), from their devel-
are unsteady and transient due to slug flow intermittency and
oped model, found the minimum stable slug lengths could reach up
irregularity. Hence, determining the mean values of these parame-
to 20D. A comprehensive analysis comparing the model solutions
ters may not be sufficient for design purposes. Accordingly, a
with the experimental data concluded that the mean slug lengths
sophisticated statistical analysis was utilized widely to provide a
were of the order of 1–2 times the minimum slug length. However,
coherent preview of the slug phenomenon. This section thoroughly
several investigators (Barnea and Taitel, 1993b, Dukler et al., 1985,
reviews the slug length, slug frequency, volume fraction distribu-
Taitel and Dukler, 1977, Nydal et al., 1992) have reported that the
tion, and slug translational velocity.
minimum stable slug length for horizontal and inclined pipes within
the range of 15D–40D. Most ancient studies concluded that the slug
3.1. Slug length length is fairly insensitive to the gas and liquid flow rates and sub-
stantially depends on the pipe diameter. However, (Wang et al.,
The entire proposed slug flow models acquired the determina- 2007) argued that the mean slug liquid length at the higher mixture
tion of the slug length, LS or frequency; mS which indicates that they velocities is comparatively insensitive to the gas–liquid flow rate. At
are the key characterization parameters. According to Perez 2008, the lower mixture velocity, the mean liquid slug length is suscepti-
the slug length estimation is challenging compared to the other ble to the gas–liquid flow rate.
slug flow characteristics. As the slug length is strongly dependent The average values and the slug parameters’ statistical distribu-
on the pipe diameter, which restricts the application of the devel- tions were investigated by Nydal et al. (1992) in pipes with 53 mm
oped correlations on small pipe diameter in the larger field-scale and 90 mm internal diameters. Air-water was used as the primary
pipes. In long pipes instead of the short ones, (Taitel, 1987) argued fluid medium. They observed that the slug length increased with
that slug length increases due to an increase of mean slug velocity increasing gas velocity and decreased with increasing liquid veloc-
at the downstream direction. Consequently, the slugs shedding ity, which is similar to the findings (Al-Kayiem et al., 2017). Nydal
rate downstream is higher than the shedding rate upstream, which et al. (1992) observed that the lognormal function was the best
increases added liquid level to the precursor slugs. Tacitly, the the- representation of slug lengths distribution, which is similar to
oretical arguments of Taitel theoretical regarding the slug length the finding of (Woods et al., 2006, Ujang et al., 2006, Bernicot
indicate that slug frequency decreases due to bubbles coalescence. and Drouffe, 1989, Bernicot and Drouffe, 1991, Dhulesia et al.,
However, due to a deficiency of relevant experimental data, he 1991, Dhulesia et al., 1993). While (Conte et al., 2017, Al-Kayiem
failed to corroborate his findings. Accordingly, his model was et al., 2017) found that the slug length was normally distributed,
unable to predict the experimental data of the Prudhoe Bay field. they claimed that slug length distributions are discrepant with
Later, (Shea et al., 2004) investigated the influence of pipe’s length (Nydal et al., 1992, Woods et al., 2006, Ujang et al., 2006) due to
(from the entry) in the slug frequency and proposed a first correla- the difference in the generated slug lengths as well as the pipe
tion that considered the pipe’s entry length as an input prediction diameter. Moreover, for large pipe (ID = 189 mm), (Kjølaas et al.,
parameter. Indeed, the distributions of slug length are evolving sig- 2020) reported that the effect of water cut on the slug length dis-
nificantly as the slugs travel downstream. Thus, the measured slug tribution is significant; however, they could not establish any rela-
length and frequency at various measuring sections vary along the tion regarding this complex phenomenon. On the contrary,
pipe (Arabi et al., 2020). Recently, in a large scale experiment with (Dhulesia et al., 1993) suggested that the inverse Gaussian distri-
a pipe of 189 mm diameter and 766 m long, (Kjølaas et al., 2020) bution is the best fit for their experimental measurements. Table 1
found that at high gas–liquid flow rates, pairs of slugs sometimes presents an overview of the slug length from the previous studies.
4
Table 1
Previous studies concern about the slug length.
Author Diameter jL (m/s) jG (m/s) Mean Comments and findings

(mm) Slug
Length
(Dukler and 38 – – 12D–30D High aeration of the slug front caused difficulty in the determination of the arrival time
Hubbard, 1975)
(Nicholson et al., 26 – – 30D Air-light oil
1978) 50
(Gregory et al., 25.8 0.03– 0.088– 30D Air-light oil fluids and some slug length reached up to 37.5 D
1978) 51.2 2.316 15.376
(Barnea and 25.4 0.01–1 0.1–10 32D the gas within the developed liquid slug behaves as dispersed bubbles
Brauner, 1985)
(Andreussi and 50 & 90 22D The mixture velocities of the air and water phases are within the range between ̴5 to ̴14 m/s
Bendiksen,
1989)
(Scott and Kouba, 609.6 0.5– 2.44– 3D–11D Slug length at low superficial gas and liquid velocities are dominated by pipeline, while at high
1990) 7.01 13.14 superficial gas velocities, increased aeration tends to increase slug length.
(Nydal et al., 53 & 90 0.6–3.5 0.5–20 15D–20D Length increases with increasing gas velocity and decreases with increasing liquid velocity
1992) 12D–16D
(Barnea and Taitel, 50 0.01 0.25 4D–14D The slug length distribution for fully developed flow follow the log-normal function
1993b) & & &
0.75 0.75 4D–24D
(Grenier et al., 53 0.35– jG < 2 m/ 20D The mean slug liquid was found to be 20D when jm > 1.2 m/s
1997) 1.6 s
(Al-Lababidi, 50 0.3– 0.6– 16D–17D Slug length measured using a non-invasive ultrasonic method and validated using conductivity
2006) 1.03 m/ 3.15 m/s measurement
s
(Woods et al., 76.3 0.6–1.2 2– 4 18D The mean length at the outlet pipe section is insensitive to the superficial liquid velocity and
2006) & & weakly sensitive to gas velocity
95 1 16D
(Ujang et al., 2006) 78 0–1.5 2–16 25D–42D The minimum and maximum slug length decreased with increasing the liquid flow rate
(Wang et al., 50 0.16– 1–20 15D–40D At higher jm velocities, the mean slug length is insensitive to gas and liquid velocity. while for low
2007) 1.5 jm velocities, the slug length decreased then increased with increasing the jm
(Kadri et al., 2010) 69 0.05– 0.2–8 16D–32D Air or Sulfur hexafluoride gas and oil were used. Slight changes in the liquid velocities can cause a
0.5 large difference in the final slug length
(Al-Kayiem et al., 74 0.7– 0.7–3.84 3D–15D the slug liquid length was observed to be normally distributed
2017) 1.33
(Eyo and Lao, 357 0.45– 0.15– ~20D– The average slug length was increased up to 1.5 times due to the eccentricity effect
2019) 3.49 2.77 75D
Few empirical correlations have been developed to predict the length are critical issues in developing the slug flow models. As
slug length. For larger diameter, (Brill, 1981) and (Scott, 1989) the pressure drop is strongly frequency-dependent, precise deter-
developed correlations that predict liquid slug length as follow: minations of the slug frequency would enhance the pressure drop
Brill et al. (1981) correlation prediction. The slug frequency model is the backbone for develop-
0:5 ing ancillary correlations, which retain the essential physics.
lnð3:2808 Ls Þ ¼ 2:663 þ 5:441 ½lnð39:37dÞ þ 0:059 (Gregory and Scott, 1969) defined the slug frequency as the num-
lnð3:2808jm Þ ð1Þ ber of slug units passed through a fixed point in the pipe over a
specific time.
While (Scott, 1989) proposed the following correlation to deter- A flush-mounted piezoelectric pressure transducer near the
mine the slug length: pipe outlet was utilized to measure the slug frequency. In this
lnð3:2808 Ls Þ ¼ 25:413 þ 28:4949 ½lnð39:37 DÞ
0:1
ð2Þ method, the pressure pulse indicates the slug passage described
by (Lin and Hanratty, 1986, Fan et al., 1993, Woods and
Based on data taken at Prudhoe Bay oil field, (Gordon and Hanratty, 1999). A conductance probe was used by (Eyo and Lao,
Fairhurst, 1988) proposed an empirical correlation for the slug 2019, Al-Lababidi, 2006, Fossa et al., 2003, Ujang, 2003) to measure
length for the pipe of 1200 , 1600 and 2000 internal diameter: the slug frequency. However, (Woods and Hanratty, 1999) illus-
0:1 trated a comparison of time series traces for the pressure captured
lnðLs =0:3048Þ ¼ 25:413 þ 28:4949 ½lnðD=0:0254Þ ð3Þ
using the pressure transducer and liquid hold-up captured using a
However, Nydal et al. (1992) proposed the following correlation conductance probe at jG = 5.5/m/s. Their comparison demonstrated
to predict the slug length in a small diameter horizontal pipe. the merit of using the pressure traces at high superficial gas veloc-
ities where it is difficult to distinguish between the aerated slug
Ls ¼ 15 D ð4Þ
and high amplitude waves. Other measuring techniques were also
used to determine the slug frequency, such as the wire mesh sen-
3.2. Slug frequency sor (Antunes et al., 2014), capacitance sensor (Al-Safran, 2009), and
high-speed camera (Al-Kayiem et al., 2017).
Slug frequencyðmS Þ, among other slug characteristics, is a crucial Due to the slug flow regime’s fickle nature, the slug frequency is
parameter that causes several operational issues such as pipeline the least reliable assessed parameter among the other slug charac-
structural instability, flooding of separators, and wellhead pressure teristics as it reflects the flow intermittency. Also, most of the
fluctuation. Besides, it was observed that high slug frequency has a researchers have demonstrated their slug frequency outcomes as
severe corrosion impact on the pipe (Hill et al., 1996). On the other a function of liquid and gas superficial velocities or in the form
hand, predicting the distributions of the slug frequency and slug of unpretentious correlations that are confined to the conditions
5
that were first developed. Therefore, few phenomenological mod- liquid velocity, pipe diameter, and slip factor. His correlation was
els have been established to determine the slug frequency (Taitel validated against experimental data (Woods et al., 2006) with a
and Dukler, 1977) developed a model to predict the slug frequency satisfactory agreement. The probability statements and distribu-
based on the slug formation process inside a horizontal pipe. The tion could be obtained utilizing the outcomes of the proposed cor-
inverse of triggered time required by the dropped liquid film to relation as an input.
rebuild and form a new slug was considered the slug frequency. (Al-Safran, 2016) concluded the pipe length is insignificant for
The rebuild process of the liquid film and slug formation is gov- slug frequency except in certain flow circumstances, particularly
erned by the following momentum and continuity Eqs. (5) and (6): at high gas flow rate with a low liquid flow rate and vice versa.
Also, (Al-Safran, 2016) recommended a thorough investigation in
@h @ v L @v L sL SL si Si 1 @P a long, high-pressure pipeline that conveys essential fluids to
g cos h þ þ uL ¼ þ þ g sin h ð5Þ
@x @t @x qL AL qL AL qL @x investigate these observations. Besides, he described that the slug
initiation mechanism is a major contributing factor for slug fre-
@h @h AL @ v L quency randomness as the developed models are considered a solo
þ vL þ 0 ¼0 ð6Þ
@t @x AL @x initiation mechanism (wave growth or coalescence). While mostly
a combination of these mechanisms exists within the same flow
Abdulkadir et al. (2016) explained that the slug length and its conditions. Woods et al. (2006) and Dinaryanto et al. (2017)
frequency are strongly interrelated slug parameters. The prediction reported that the slug frequency primarily depends on the liquid
of the slug frequency based on slug length can be obtained using superficial velocity and it would increase when the liquid superfi-
Eq. (7). However, Eq. (7) is valid for constant slug length only, cial velocity increased. As the slugs move toward the pipe outlet,
and since the distribution of the slug length is evolved substan- the slug frequency decreased due to coalescence or collapse effects
tially as the slug travel along the pipe, the usage of Eq. (7) is lim- (Dinaryanto et al., 2017).
ited. Although Eq. () illustrates that the slug length and Furthermore, (Dinaryanto et al., 2017) added that significant
frequency are firmly related, it is still arguably enough to report endeavors are required to resolve the developed slug frequency
only one of these parameters (Kjølaas et al., 2020). correlations’ discrepancies and inherent uncertainty. On the other
hand, (Al-Kayiem et al., 2017) concluded that the slug frequency at
jB
mS ¼ ð7Þ constant superficial gas velocity increased with increasing superfi-
LS þ LB
cial liquid velocity and decreased at constant superficial liquid
However, several correlations have been developed to predict velocity with increasing the superficial gas velocity. These observa-
the slug frequency in horizontal pipes, as shown in Table 2. tions are similar to the findings of (Abdulkadir et al., 2016, Fan
The aforementioned studies demonstrated that the applicability et al., 1993, Manolis et al., 1995, Fossa et al., 2003). On the contrary,
of those correlations is limited and restricted for specific condi- (Wang et al., 2020) used the fast Fourier transformation method to
tions. However, some models are too general. (Taitel and Dukler, determine the slug frequency. He found that the superficial liquid
1977), which is the only valid model for broad flow conditions, velocity has an immense effect on the slug frequency distribution,
including inclined flow. On the other hand, some correlations are while the impact of the superficial gas velocity on the slug fre-
straightforward yet don’t involve the entire affecting variables of quency is insignificant. Additionally, (Wang et al., 2020) outcomes
the slug frequency; thus, they are less precise. Hill and Wood were in good agreement with (Dinaryanto et al., 2017).
(1990, 1994) and (Shea et al., 2004)are the third correlations group As reported by (Arabi et al., 2020), each correlation was devel-
that is applicable for a broad spectrum of conditions. However, oped based on the available data, which interprets the significant
their usage is tedious as they comprise independent parameters. discrepancy of the predictions between different correlations.
Later, (Shea et al., 2004) demonstrated that the (Taitel and Heretofore, there is no unanimous correlation valid for the entire
Dukler, 1977) model over-predicting the slug frequency notably operational conditions. Therefore, (Arabi et al., 2020) determined
when applied for large-scale field data, with an approximate error the slug frequency based on three methods, which namely are:
of 2000%. Simultaneously, a good validation for (Taitel and Dukler, counting method (Hubbard, 1965), differential pressure method
1977) model was achieved by (Zabaras, 1999) with a tiny bunch of (Wilkens and Thomas, 2008), and Power Spectral Density (PSD)
laboratory data for the horizontal and inclined flow. Besides, method (Bertola, 2003, Lin et al., 2013). He found that the PSD is
(Zabaras, 1999) mentioned that the difficulties of developing a the most suitable technique for computing the slug frequency.
sophisticated numerical solution are attributed to the high compu- Based on a collected experimental data, (Arabi et al., 2020) pro-
tation time required to calculate a single slug frequency. Also, he posed a new correlation that considers the intermittent flow sub-
mentioned that the expensive computational cost is the major regimes. Best predictions of the experimental data were obtained
drawback of the Taitel and Dukler model. Later, (Taitel and using this correlation. However, (Kim et al., 2020) explained that
Dukler, 1977) validated their model with short slugs data that an ancillary experimental data is still required to identify the
merged with elongated slugs. Accordingly, (Taitel, 1987) justified appropriate correlations for the slug characteristics. Morover,
the discrepancy of their model is ought to the model deficiency (Kim et al., 2020) recommended that the usage of the artificial
in capturing the slug frequency at the pipe inlet, which conse- intelligence method such as machine learning could be a suitable
quently underrates slug frequency. Furthermore, (Taitel, 1987) route to eliminate the subjective utilization of the previous empir-
described that their numerical solution acquired a few seconds to ical correlations as these correlations required the experimental
run on an old computer, which contradicting with (Zabaras, slug characteristics known in advance.
1999) finding as reported by (Al-Safran, 2009). Based on the afore-
mentioned discussion, the prominent Taitel and Dukler model is 3.3. Slug velocity
applicable and more accurate for the cases of not fully developed
flow where the slug frequency is relatively high and the slug length The normal slugs propagation and development are substan-
is short (Al-Safran, 2009). Furthermore, the entire presented corre- tially depending on the shedding rate at the back and the pickup
lations and models are deterministic and none of them is capable rate of the liquid at the slug front. However, there is a strong rela-
to determine the related probability of the predicted slug fre- tionship between the shedding and the pickup rates, as the pickup
quency (Al-Safran, 2016). Therefore, (Al-Safran, 2016) proposed a is the process of collecting the liquid left behind the preceding slug.
new mean slug frequency correlation as a function of superficial The shedding process is caused by distributing the slug tail velocity
6
Table 2
The developed slug frequency correlations.
Authors Pipe Fluids Slug frequency Correlation

Diameter
(mm)
(Gregory and Scott, 19 Co2- water h i1:2
mS ¼ 0:0226 jL
gD
19:75
jm
þ jm (8)
1969)
(Greskovich and 38–152.4 air–water h i1:2
mS ¼ 0:0226 jjL 2:02
D þ Fr S (9)
Shrier, 1971) m
FrS ¼ pjmffiffiffiffiffi (10)$1

gD
(Heywood and 42 air–water h i1:02
mS ¼ 0:0434 jjL 2:02 D þ Fr S (11)
Richardson, 1979) m
(Fetter, 1988) 82 air–water- mS ¼ 0:0175Fr1:37

Fet (12)
oil
h i
jL 21:3þj2m
FrFet ¼ gD jm
(13)$1
h i
(Hill and Wood, 1990) 75 air- mS ¼ 0:275
3600
jL þjG
D :102:68ð1ae Þ (14)
200–400 kerosene
air–water
(Nydal et al., 1992) 52.9 and 90 air–water mS ¼ 0:088 2
gD ½jl þ 1:5 (15)
(Jepson and Taylor, 306 air–water mS ¼ jDL ½0:00759jm þ 0:01(16)
1993)
0
(Hill and Wood, 1994) 75 air- mS D
0 0
jm ¼ 24:73 þ 0:00766e 9:9121HLe

þ 24:72eð0:20524HLe Þ (17)
150.4–400 kerosene
air–water
0
mS D
jm ¼ mjS D ð1 0:05jG ÞD0:3 (18)
m
0
HLe ¼ HLe ð1 0:068
j Þ (19) L
(Manolis et al., 1995) – air–water mS ¼ 0:037Fr1:8

Man (20)
2
jL jm;min þj2m
FrMan ¼ gD jm
(21)a
(Zabaras, 1999) 25–203 air–water h i1:2 h i

mS ¼ 0:0226 jL
gD
19:75
jm þ jm 0:836 þ 2:75 sin0:25 h (22)
air-oil
(Shea et al., 2004) – – mS ¼ 0:47j L
(23)
0:75
D1:2 L0:55 P

(Al-Safran, 2009) 50.8 air-oil LnðmS Þ ¼ 0:8 þ 1:53LnðjL Þ þ 0:27 jL
34:1ðDÞ(24)
jm
(Gokcal et al., 2010) 50.8 air-oil tS ¼ 2:623 N0:612 ðdÞ(26) jL
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f
d
3=2
qL ðqL qG Þg
Nf ¼ lL (27)
(Schulkes, 2011) – – tS ¼ wðaÞ UðNRe;L Þ Hðh; NFr Þ(28)
( pffiffiffiffiffi
12:1N 0:37 1 þ N2Fr sin h jhjFor jhj 0:17
wðaÞ ¼ 0:016 / ð2 þ 3aÞU N Re;L ¼ Re;L ForN Re;L < 4000 Hðh; N Þ ¼
Fr
1ForN Re;L 4000 1:8
NFr ð0:6 þ 2h h2 ÞFor jhj > 0:17
(
(Zhao et al., 2013) 26 and 74 air-oil
tS 10:84N 0:38 4000
Re;L ForN Re;g
wðaÞ ¼ (29)
6:40N 0:141
Re;L > 4000
ForN Re;g
(Picchi et al., 2015) 73.66 air- CMCb h i0:88 0:07
mS jL 32:2014
¼ 0:0448 gD j þ jm n2:85 Re
Ren
(30)
(water) m w
(Al-Safran, 2016) 25–203 air-oil mS ¼ EXPð1:51 17:04ðDH Þ þ 0:77LnðjL Þ 0:181LnðsÞ(31)

air-
kerosene Where S is the slip factor
air–water
Freon-
water
N2-Disel
a
Frs, FrFet, and FrMan denotes the Froude number of the slug frequency, Fetter, and Manolios Froude number.
b
CMC denotes CarboxyMethyl Cellulose solution.
profile near the slug liquid (Dukler and Hubbard, 1975, Nicholson
et al., 1978). Due to friction, the liquid velocity increased in the
radial direction toward the pipe centerline and decreased near
the pipe wall. Consequently, the amount of the liquid from the slug
body moves with small velocity backward and forms a liquid film
behind the slug. This liquid film moves with a velocity similar to
the slug translational velocity, called the interface velocity, ji.
Therefore, the shedding rate is assessing the slug translational
velocity (Taitel, 1987). Fig. 2 illustrates the tail slug velocity profile
Fig. 2. The shedding mechanism.
behind the liquid slug.
Slug velocity plays a significant role in understanding the
gas–liquid slug flow because it estimates slug kinetic energy, conducted experimental works have focused on the velocity of a
consider the fluctuation of the pressure pulse, and the effect of single elongated bubble rather than a full slug unit’s motion.
the flow rate variations. Unfortunately, most of the previously Besides, most of the previous correlations have been developed
7
based on ducts with an internal diameter not greater than 0.05 m velocity and the superficial liquid velocity. Thirdly, aerated slugs
(Bertola, 2002). For decades, the propagation of long bubbles in the generated small bubbles that coalesce at the bubble nose and conse-
pipe has been thoroughly investigated. Previous researchers, e.g. quently affect the translational bubble velocity. However, (Hale,
(Dumitrescu, 1943, Davies and Taylor, 1950, Collins et al., 1978, 2001) has addressed these factors’ effect on the translational veloc-
Bendiksen, 1984, Bendiksen, 1985) have computed the bubble ity by conducting a ‘‘gassy pushout” experiment. He realized that the
velocity through 3D liquid flow around the bubble nose. While distribution parameter value C0 was increased with increasing the
the other developed correlations for horizontal gas–liquid slug slug liquid’s aeration level, as shown in Figs. 3 and 4. Furthermore,
flow have determined the bubble velocity at its tail, these correla- the distribution parameter’s augment was attributed to several fac-
tions can be classified into two groups. The first group predicts the tors, such as the strong relationship between the distribution
slug tail velocities based on the mean fluid velocity (Dukler and parameter and the slug liquid velocity profile. In addition to the con-
Hubbard, 1975, Nydal et al., 1992). In contrast, another group tribution of the drift velocity to translational slug velocity and the
determines the slug tail velocities based on Froude number and de-aeration of the liquid film (Hale, 2001). Therefore, a modified cor-
the drift velocity (Nicholson et al., 1978, Ferre, 1978, Bendiksen, relation was introduced by (Hale, 2001) as follow:
1984, Fabre and Liné, 1992). Generally, the correlations of both
jB C o ð1 eGS Þ jD
groups can be summarized using the developed correlation by ¼1þC ¼ þ ð33Þ
jm ð1 C o eGS Þ jm ð1 C o eGS Þ
(Nicklin, 1962) as follow:
where eGS is the gas holdup at the centerline position
jB ¼ C o jm þ C 1 jD ð32Þ
A viscosity number was added to the distribution parameter for
where jB, Co, jm, and jD are the slug velocity, the distribution param- a highly viscous fluid to comprise the effect of the liquid viscosity
eter, mixture velocity, and drift velocity. The effect of pipe diameter on the translational slug velocity. Viscosity number was success-
on the distribution parameters of the mixture velocity and drift fully utilized by many authors (Al-Safran et al., 2011b, Al-safran
velocity is negligible. et al., 2013b, Baba et al., 2018, Baba et al., 2019) to characterize
The distribution parameter Co in Eq. (32) is derived by assuming highly viscous slug flow in the pipelines, which yield the slug
that the slug bubble velocity is similar to the maximum local liquid translation velocity as follow:
velocity jmax at the slug nose (Nicklin, 1962, Bendiksen, 1984,
jB ¼ C o þ N l jm þ C 1 jD ð34Þ
Shemer, 1987, Polonsky et al., 1999). Therefore, the distribution
parameter is determined as the ratio between the maximum to where
the mean liquid velocities inside the pipe, which is approximately
equal to 1.2 for fully developed turbulent flow and 2.0 for laminar jm l L
Nl ¼ ð35Þ
flow (Van Hout et al., 2002). Recently, (Bendiksen et al., 2018)indi- gD2 ðqL qG Þ
cated the importance of considering the pipe diameter on the slug
On the other hand, (Ruder et al., 1989) defined slug front as a
tail velocity computation. Accordingly, he introduced the slug tail
hydraulic jump and utilized this definition to introduce the neces-
velocity as a function of pipe diameter, inclination angle, gas–liq-
sary criteria for the existence of non-aerated stable slug in the pipe.
uid properties, mixture flowrates, and the associated non-
(Ruder et al., 1989) derived these criteria based on the inviscid
dimensional numbers. These non-dimensional numbers are Rey-
steady-state model (Benjamin, 1968) and formulate it as follow:
nolds number (Re), Froude number (Fr), Archimedes number (Ar),
and Eötvös number (Eo) (Wan et al., 2019), which they are depend- ðjF jLF Þ 4ALF
¼ 0:54 ð36Þ
ing on fluid properties and flow conditions. Various endeavors ðgDÞ1=2 pD2
have been conducted to evaluate the distribution parameter, Co
where jLF and ALF are the velocity and area of the liquid film ahead of
and the drift constant, C1, as tabulated in Table 3.
the slug unit.
Compared to the vertical pipe configurations, fewer works have
While (Bendiksen and Espedal, 1992) estimated the slug front
been conducted to investigate the motion of elongated bubbles in
velocity based on the mass balance across the slug front, presum-
the horizontal pipe. Investigations scarcity could be attributed to
ing that there is no droplet entrainment on the gas phase. In their
the complexity of modeling an asymmetric bubble shape. Conse-
calculation, the expression introduced by (Bendiksen, 1984) for the
quently, the drift velocity was avoided or purposely ignored by most
translational velocity was used to determine the slug tail velocity.
of the previous studies due to the trivial gravity effect in the horizon-
Then, the slug front velocity was estimated by:
tal direction (Wallis, 1969, Dukler and Hubbard, 1975). However,
later investigations (Nicholson et al., 1978, Bendiksen, 1984, eGF eGS jjGFG
Kouba and Jepson, 1990) revealed that the gravity effect induced jF ¼ jGF ð37Þ
eGF eGS
drift caused by the different bubble nose elevations. These discrep-
ancies caused ample distribution of the parameters Co (0.95–1.54) where jGF, jG, eGF,eGS are gas velocity on the region above the liquid
and the drift constant, C1 (0–0.6). In the same context, Bendiksen film, gas velocity, superficial gas velocity, the void fraction of the
1984, discovered the plot of the translational velocity versus mix- region above the liquid film in front of the slug, and gas void fraction,
ture velocity is non-linear as it bends slightly upwards. Accordingly, respectively. By contrast, the slug front velocity jF was assumed like
it was observed the drift velocity determined after fitting the exper- the translational velocity of a head slug. This assumption was formu-
imental data was linearly different from the calculated drift velocity lated based on the condition that the slug bubbles not crumbling as
before plotting the linear fit. Hence, three potential reasons were they flow inside the pipe (Barnea and Taitel, 1993b).
introduced by (Nydal et al., 1992) to justify the differences between Moreover, the investigations of (Ruder et al., 1989) were
experimental distribution parameter C0 and the measured value extended by (Woods and Hanratty, 1996) to determine the liquid
obtained by the mean to the centerline velocity ratio. Firstly, bubbles velocity with the slug body considering the slip between the
that appear in the slug flow create a difference between the slug liq- gas–liquid phases. The superficial liquid velocity was determined
uid velocity in front of the bubble and the centerline velocity of pure using Eq. (38):
liquid, which consequently alter the slug tail velocity profile at the jm
same Reynolds number (Andreussi et al., 1991). Secondly, as the slip jL ¼ ð38Þ
½1 þ ðS 1ÞeGS
occurs between the gas–liquid phases, the distribution parameter
fluctuates to compensate for the variation between the mixture where the slip ratio S within the slug liquid is given by
8
Table 3
Comparison of the distribution and drift constant parameters from the previous studies.
Authors ID Co C1 jL (m/s) ReL (103) Remark

(mm)
(Nicklin, 1962) 1.2 0.35 – – 1.2 for turbulent
2 2 for laminar flow
(Gregory and Scott, – 1.35 0 – – jB ¼ C o jm
1969)
(Dukler and Hubbard, 38.1 0.021 ln(Re) + 0.022 0 – 30–400 Smooth friction factors
1975)
(Nydal et al., 1992) 52.9 1.2 0 0.6–3.5 32–184 Drift velocity was neglected
90
(Nicholson et al., 1978) 25.8 1.196 0.54 0.03– 0.1–7 Drift velocity is existing for horizontal pipe
51.2 1.128 0.4 2.13 0.2–7.8
0.06–
2.29
(Ferre, 1978) 45 1.1 0.24 05–2 22.4–90 Frcrit < 2
1.3 0 Frcrit < 2 < 8
1.02 1.61 Frcrit > 8
(Bendiksen, 1984) 24.2 1.2 0.54 0.3–5 0.6 < 10.4 FrM > Frcrit
1.05 FrM Frcrit
(Dukler et al., 1985) 38.1 1.225 0 – – Co is the 7th power law for turbulent flow
(Kouba and Jepson, 150 1.25 – 0–6 – At jm < 3 m/s, the drift velocity should be considered and Co = 2
1990) 2
(Manolis, 1995) 74 1.033 0.48 0.05– – If Frc 2.86
0.51 C 1 ¼ 0C o ¼ 1:216
(Woods and Hanratty, 95.3 1.1 0.52 0.6–1.5 – At jm > 3 m/s,
1996) C 1 ¼ 0C o ¼ 1:2
h i0:25
(Petalas and Aziz, – :031
(1:64 þ 012sinhÞReml 0:3 1 eb – –
jB ¼ 1:53 g ðqLq2qG Þr sin h
2000) L
pffiffiffiffiffiffiffiffiffiffi
(Choi et al., 2012) 76.2 1:20:2 qG =qL ð1expð18aG ÞÞ 0.01–1 – To consider the inclination angle
2
þ 0:25
1þðRe =1000Þ2 1þðRe =1000Þ2
jD ¼ A cos h þ B gqrLD2q sin h
(Kim et al., 2018) 50.8 2.11–2.2 – 0.2–1 – for the elongated bubble flow Co = 2.11–2.2 while for laminar slug
2.27 flow Co = 2.27
(Wan et al., 2019) 50 1 0.3–8.2 6.7–160 atmospheric pressure
27.9 1.04 4.2
54.8 0.95 1.35 0.2–3.5 10–234 At 1 bar
108 0.5 0.2 1.7–7.3 1.7–7.3
Fig. 3. Distribution parameter as a function of mixture velocity - Low-Pressure Rig (Hale, 2001).
jG Thus, the translational slug velocities can be estimated using Eq.

S¼ ð39Þ
jL (41)
Therefore, the distribution parameter for the case of the slip jB ¼ C o;s jm þ jD ð41Þ
velocity within the slug liquid was obtained by
Later, (Bendiksen et al., 2018) explained that the slug transla-
Co
C o;s ¼ ð40Þ tional velocity is too independent of slug length LS/D > 5, while the
½1 þ ðS 1ÞeGS slug front is relatively independent slug length LS/D > 5. Also,
9
Fig. 4. Distribution parameter as a function of mixture velocity – WASP (Hale, 2001).
(Bendiksen et al., 2018) mentioned that the slug front velocity might Raithby, 1995a, De Henau and Raithby, 1995b, Barnea and Taitel,
vary significantly along the pipe if only one slug flows inside the pipe. 1993b, Nydal and Banerjee, 1996, Al-safran et al., 2004, De
Furthermore, (Wan et al., 2019) explained the difficulties behind Leebeeck and Nydal, 2010) simulated the slug flow using a two-
modeling slug front could be attributed to its chaotic, turbulent nat- fluid model employing slug unit and slug tracking schemes. How-
ure. A random distribution of the slug front velocity was observed ever, modeling of the slug flow initiation and the dynamic growth
with the slug translational velocity, yet the slug front value is pre- using these schemes is challenging, as they relied on the empirical
dictable when the slug translational velocity is known. However, correlations to assess the slug properties (Nieckele et al., 2013).
when the slug translational velocity is unknown, the slug velocity Therefore, a hybrid method proposed by (Renault, 2007) to lessen
became unpredictable. Therefore, (Wan et al., 2019) resolved this the dependency on the empirical correlations, but the prior need
issue through developing a statistical deterministic problem which for the statistical distributions of the slug characteristics has
allows modeling the slug front velocity as a function of mean slug remained a challenge. By contrast, the high resolution capturing
translational velocity and internal pipe diameter as follow: techniques were utilized by (Issa and Kempf, 2003, Carneiro
et al., 2011, Danielson et al., 2012) to capture the hydrodynamic
jF ¼ AJ bB þ C ð42Þ instabilities of the gas–liquid interface. The capturing of the hydro-
where dynamic instabilities and the transition from stratified to slug flow
was performed numerically by solving the two-fluid model as
A ¼ 6:7D þ 0:16; b ¼ 0:7D0:25 ; C ¼ 0:6 demonstrated by (Issa and Kempf, 2003, Carneiro et al., 2011).
However, several challenges were posed while modeling the
hydrodynamic slug flow, such as modeling the dynamic behavior
4. Numerical modeling of slug flow
of the gas–liquid interface, which separates both phases. Mean-
while, the occurrence of significant air entrainment is leading to
The slug initiation, growth, and development have been pre-
concurrent gas dispersion into the slug liquid and liquid droplets
dicted intensively using numerical simulations. However, most of
into the gas bubble. The predictions of these bubbles and droplets’
the existing numerical models required closure relationships due
size are extremely difficult owing to the intricate turbulence phe-
to the difficulty of possessing the whole flow process details (Lu,
nomena in the vicinity of the elongated gas–liquid interface. Due to
2015). In general, two simulation methods were used to model
these complex, interrelated phenomena, the current 1D-modeling
the slug flow which is: ‘‘slug-tracking” which utilizes the adapta-
was found inadequate to provide precise predictions (Mo et al.,
tion of the basic unit cell model to track the tail and front of the
2014). Fig. 5 illustrates the typical time tracking of the liquid
slug (Zheng et al., 1994, King, 1998, Manfield, 2000, Rosa et al.,
holdup in slug flow.
2015, Nydal and Banerjee, 1996, Rodrigues et al., 2008, Smith,
Moreover, for modeling the slug flow in horizontal pipes, three
2017, Naidek et al., 2017, Han et al., 2019); ‘‘slug-capturing” where
computational approaches have been identified by (Frank, 2005),
slugs growth are initially captured from stratified to slug flow tran-
which are:
sition until the fully-developed slug flow. The slug-capturing
approach is based on the transient two-fluid model equations
Frozen slug with moving wall boundaries. In which the prede-
(Issa and Kempf, 2003, Renault, 2007, Ansari and Shokri, 2011,
fined wall velocity is equal to the slug propagation velocity.
Kjølaas et al., 2013, Bonizzi and Issa, 2003, Bonizzi, 2003,
Thus, the propagation velocity, slug length, and slug frequency
Montini, 2011, Azizi et al., 2019, Ferrari et al., 2019). For long pipes,
must be known before executing the simulation.
One-dimensional modeling of two-phase flows was broadly used
Three-Dimensional (3D) simulation in a short pipe with peri-
on diverse codes comprising different modeling formulations. Most
odic boundary conditions. In which, the pressure losses due to
of the prediction Methods of slug dynamics are strongly dependent
wall shear forces should be compensated by defining a driving
on the flow transition criteria and the predefined initial conditions
pressure force at the inlet boundary condition.
and the slug initiation assumptions. The studies of (De Henau and
10
Ghiaasiaan, 2006, Qian and Lawal, 2006, Kashid et al., 2008,

Carlson et al., 2008, Fukagata et al., 2007). Fig. 6 illustrates a com-
parison between the experimental and numerical slug flow in a
horizontal pipe, in which the interface and slug flow morphology
were predicted with high accuracy. From Fig. 6, the dispersed bub-
bles and the water droplet in the gas phase were not reproduced by
the developed VOF model as it requires dense mesh for this small-
est length scale. Therefore, to capture the infinitesimal details of
the complex flow, the two-fluid model became the best alternate
in term of the computational cost (Godino et al., 2020).
There are other methods, such as LS (Fukagata et al., 2007),
Lattice-Boltzmann (Yu et al., 2007), or Phase-field (He and Kasagi,
2008). As reported by (Horgue et al., 2012), the thin liquid film in
the slug flow was not captured successfully in the previous studies
using diffusive-interface methods as they induce significant errors
in the prediction of flow properties. For accurate prediction, these
methods required ultra-dense mesh with long computational time,
Fig. 5. Real-time tracking of liquid holdup for the slug flow in a horizontal pipe (Issa which restricts their applications at large scale model. However,
and Kempf, 2003). (Horgue et al., 2012) concluded that suitable parametrization of
the VOF method could precisely capture flow properties compris-
ing thin liquid films with high accuracy and less computational
Three-Dimensional (3D) Transient simulation in a long pie with cost. The process of bubble generation is shown in Fig. 7.
inlet/outlet boundary conditions. The typical inlet boundary The aforementioned models provide comparatively precise and
condition of the experiment is acquired to simulate the slug for- reliable outcomes for either dispersed or segregated flows. How-
mation and predict the slug characteristic numerically. Initially, ever, in real-world applications, most of the multi-phase flows
the flow morphology, shape, interfacial area, and the free sur- are a combination of a successive two flow regime. For instance,
face that separates gas–liquid phases are unknown parameters. slug flow is a simultaneous amalgamation of aerated slug liquid
Thus, the transient 3D simulation with inlet/outlet boundary (dispersed flow) and elongated bubbles (segregated flow). There-
conditions is a suitable technique to compute the slug charac- fore, the previous models are impractical methods of simulating
teristics (Mohmmed et al., 2018). the two-phase slug flow (Akhlaghi et al., 2019). Accordingly, The
Multifluid-VOF model, which is a combination of two-fluid and
Recently, after the computer revolution, light has been shed on VOF approaches, has been developed by (Cerne et al., 2001). Ini-
the usage of Computational Fluid Dynamics (CFD) to investigate tially, this method was developed to simulate 2D Rayleigh–Taylor
two-phase flow using numerous approaches. All these approaches instability. Multifluid-VOF is capable of modeling a large scale of
under the umbrella of 3D transient simulation with inlet/outlet the gas–liquid interface in segregate flow regimes as well as utiliz-
boundary conditions, which are the Eulerian-Lagrangian (Hirt ing the mean values for each phase in the dispersed flow. The
et al., 1974), Eulerian–Eulerian (Ishii and Mishima, 1984), and reconstruction accuracy of interface capturing was enhanced to
the interface capturing techniques. The Eulerian-Lagrangian lessen the diffusion effects. (Yan and Che, 2010, Štrubelj and
approach does not allow the modeling of a large interface; thus, Tiselj, 2011, Denèfle et al., 2015) utilizing the Level Set method
it is frequently applied for cases where a dispersed phase exists, for tracking the interface in the Multifluid-VOF method. The
such as bubbly flow. While the Eulerian–Eulerian approach is an Multifluid-VOF model was built in the OpenFOAM Software to sim-
average time approach that utilizes a single set of equations for ulate a liquid–liquid extraction device. Also, (Denèfle et al., 2015)
each fluid separately. Whereas, the interface capturing techniques was investigated gas–liquid churn flow in the vertical pipe using
encompasses various method such as Volume of Fluid (VOF) (Hirt the developed Multifluid-VOF method in ANSYS software. The
and Nichols, 1981) and Level Set (LS) (Sussman et al., 1994). VOF results exhibited a reasonably good correlation between the
method shares a single set of equations for both gas–liquid phases. empirical and numerical mean void fraction. Also, (Parsi et al.,
It uses ancillary transport equations based on the volume fraction 2016) utilized the Eulerian–Eulerian MultiFluid VOF model to sim-
to distinguish between the phases at each cell. (Guerrero et al., ulate high flow rate air–water multiphase flow in a 76.2 mm-
2017) conducted a comparison between Eulerian and VOF models diameter pipe. Their model was successfully predicting the phase
based on modeling six experiments with an orthogonal mesh. They distributions, mean void fractions, and average void fraction time
found that the Eulerian model provides results with the least mean series. A comparison with experimental data revealed good agree-
square errors in comparison to the VOF model at low void fraction ment between the numerical and experimental results in terms of
flow. Besides, their results demonstrated that the Eulerian model is capturing different liquid structures within the flow and the void
insensitive to grid number variation. Accordingly, it requires less fraction data. Recently, (Akhlaghi et al., 2019) used the
computational time than VOF. Furthermore, their results show that Multifluid-VOF model to simulate the horizontal air–water inter-
only the VOF model can predict and identifying the flow pattern. mittent flow considering the gravity force’s effects on the flow
Therefore, VOF is a suitable method for two-phase applications field. A pipe of 44 mm diameter and 6 m length was modeled in
because it allows tracking the interface and the liquid holdup as OpenFOAM software, and the simulations were executed on Mul-
well as its capability of predicting the flow regime (Pineda-Pérez tiphaseEulerFoam solver. A detailed comparison between Multi-
et al., 2018). Due to the ultimate prediction accuracy and low com- Fluid VOF and VOF models was conducted and the numerical out-
putational cost, the VOF method has been used extensively to comes of both methods have been assessed against experimental
model the slug flow in a horizontal pipe (Mohmmed et al., 2018, outcomes, as shown in Fig. 8. An excellent agreement between
Al-Hashimy et al., 2016, Mohmmed, 2016, Ghorai and Nigam, the Multifluid-VOF and the experimental results was achieved.
2006, Lu et al., 2007, Andrianto et al., 2016, Horgue et al., 2012, Furthermore, (Akhlaghi et al., 2019) concluded that the simulation
Frank, 2005, López et al., 2016, Pineda-Pérez et al., 2018, Al- time is the only disadvantage of the Multifluid-VOF over VOF as it
Kayiem et al., Gupta et al., 2009, Taha and Cui, 2004, Akbar and entails 14 times the simulation execution of VOF.
11
Fig. 6. Comparison between the experimental and numerical water volume fraction (Mohmmed et al., 2020).
Fig. 7. VOF simulation in a microchannel using the HRIC Scheme (a) the process of bubble generation (b) volume fraction and pressure distribution (Horgue et al., 2012).
On the other hand, (Lakehal et al., 2002) discussed the trends of they introduced a novel approach referred to as the boundary fitting
various prediction methods related to the numerical simulation of method. Later, (Lakehal et al., 2012) used LS Techniques combined
multiphase flows based on the one-fluid formalism. Their discussion with the Very-Large Eddy Simulation (V-LES) approach to simulate
covers broad spectrums of interface tracking algorithms coupled the transition from stratified to slug flow. They utilized this model
with one-fluid formalism. These algorithms are based on solving a to reproduce WASP experimental data along with the droplet
single set of transport equations for the entire computational entrainment in a stratified flow. Their 3D model provides clear
domain, which considers gas–liquid phases as a single fluid with dif- insights about the flow details; in particular, the slug formation
ferent material properties. This study concluded that LS method has and frequency were predicted precisely. Their model also provides
limited applicability due to its inherent inability to conserve mass. promising results about the droplet entrainment, including the liga-
While the VOF approach coupled with the Front Tracking (FT) ment formation and sheet break up into droplets. However, (Lakehal
scheme is more accurate than LSs, their extension to an extremely et al., 2012) commented that the grid refinement is still an
complex 3D multiphase flow poses serious challenges. Therefore, unresolved issue, particularly for turbulent flow cases where the
12
Fig. 8. Void fraction distribution of slug and plug flow under various superficial liquid and gas velocities (Akhlaghi et al., 2019).
generated droplet and entrained bubbles are massive. Therefore, to determine the slug birth rate along with the slug initiation in mul-
(Mo et al., 2014) averaged down the 3D method to 2D using a unique tiphase fluid flow is introduced by (Lawrence and Hu, 2020). They
Quasi-3D approach to obtain a compromise solution in terms of cost determined the slug birth rate based on the difference between
and accuracy. This model was originally developed by (Laux et al., the slug front and slug tail velocities.
2007), and it is based on a multi-fluid multi-field formulation com-
bined with construction and tracking of the large-scale interfaces
(LSIs). In which the equations are derived based on ensemble and 5. Factors affecting the slug two-phase flow characteristics
volume averaging of the Navier-Stokes equations. (Mo et al., 2014)
utilized this model to predict the experimental results of the slug ini- 5.1. Drag and frictional effect
tiation and the subsequent evolution of hydrodynamic slug in a hor-
izontal pipe. These numerical investigations were performed based According to (Toms, 1948), adding a tiny concentration of high
on experiments carried out by (Ujang et al., 2006). Fig. 9 illustrates molecular weight polymer to any solvents is substantially reducing
snapshots of the slug evolutions in the horizontal pipe. Parts of the the frictional pressure drop in turbulent flows. Which conse-
(Laux et al., 2007) model concept are also detailed in a patent for quently enhance the pipeline capacity and production rate. Slug
steady-state flows (Johansen, 2019). Recently, (Schmelter et al., two-phase flow has various distinguished features that must be
2021) investigated the influence of the interface instability imposed investigated in each situation. Unlike the other two-phase flow
by the inlet velocity perturbation using the VOF model. Their simu- phenomena, the axial pressure gradient is highly undesirable as
lation results provide a first systematic guideline regarding the it creates considerable energy consumption per unit volume of liq-
impact of inlet perturbations slug growth and initiation in the pipe. uid (Mowla and Naderi, 2006). The turbulence effect in the gas–liq-
However, they recommended that future work should address the uid two-phases is higher than in the single-phase (Liu, 2013),
correlations related to unstable slug and saturated slug flow with especially in oil and gas pipelines. Thus, the prospective reduction
registered uncertainties of measurement. Moreover, a patent aims of drag in two-phase turbulent flows is larger than single-phase
13
Fig. 9. Snapshots of Q3D results showing the evolution of slugs in a 30 m long pipe (Mo et al., 2014).
flows. However, few attempts have been conducted to study the

drag reduction on gas–liquid two-phase flow (Wang et al., 2019).
In the first study of two-phase drag reduction, the interfacial waves
were suppressed by a macro-molecular-polymeric drag reducer
(Oliver and Hoon, 1968). Previous studies (Kang and Jepson,
1999, Kang and Jepson, 2000, Daas et al., 2000) discovered for a
high polymer concentration of 10 ppm to 50 ppm, the drag reduc-
tion amount varied from 30% to 50% for gas–liquid two-phase flow.
Also, for high superficial gas velocity, the drag reduction decreased
from 50% to 30%. Besides, (Kang et al., 1999) found that adding
50 ppm drag reducing agents cause a flow regime transition for
slug two-phase flow to stratified flow. More studies were reviewed
thoroughly by (Al-Sarkhi, 2010).
On the interval between every two successive slugs, frictional
losses occurred in both of the slug unit components. Therefore,
an excessive pressure loss in the slug liquid occurred, which highly
affect the slug characteristics. The rate of the pressure losses due to Fig. 10. The effect of polymer concentration on the liquid holdup at various
friction has proportional relation with the gas content inside the superficial liquid velocity (Soleimani et al., 2002).
slug liquid and the slug velocity. A longer slug liquid has greater
accompanied pressure loss comparing to the short slug liquid.
flow. Additional studies concern about the drag reduction effect on
However, shorter slug liquid usually has high slug frequency,
two-phase flow behavior can be found in (Pouranfard et al., 2015,
which causes multiple significant magnitudes of the total pressure
Alsurakji et al., 2018, Nesyn et al., 2018). All these studies
loss. Using the drag reduction agents can reduce the total pressure
described various crucial factors on the drag reduction rate of
losses through any of its components, consequently decreasing the
gas–liquid turbulent flow. However, most of these findings are
slug frequency (Daas and Bleyle, 2006).
based on experimental observations, which address the two-
On the other hand, it was observed that critical liquid holdup,
phase drag-reducing flow’s general behavior. Whereas the local
hL/D, and JL are increasing with increasing the polymer concentra-
characteristics have not been extensively investigated as reported
tion for the initiation of slugs and roll waves, as shown in Fig. 10. At
by (Wang et al., 2019). Deficiency in understanding the local
the same time, the interfacial friction coefficient decreased with
behavior of the two-phase drag-reducing flow induced by polymer
increasing the polymer concentration, which might be significantly
additives bounds the understanding of the drag reduction mecha-
large in the pseudo-slug regime. Due to the reduction of the inter-
nism in two-phase flow.
facial friction fi, a noticeable increase of hL/D was observed. Also, a
significant decrease in the frictional pressure drop was observed
(Soleimani et al., 2002). As the polymers suppress the turbulence 5.2. Liquid viscosity effect
effect of the slugs and damp waves at the interface. (Soleimani
et al., 2002) suggested that slug laminarization is associated with For decades, various models have been developed to predict the
the increase of the transitional slug velocity. slug flow behavior. However, modern defiance has emerged in the
It was discovered by (Al-Sarkhi and Hanratty, 2001, Soleimani oil and gas industry, where high viscosity liquid and gas flows are
et al., 2002, Al-Sarkhi and Soleimani, 2004) from their experimen- considered. As reported in the previous studies, the Liquid viscosi-
tal studies; polymer drag-reducing agents provide ultimate drag ties were significantly affecting the two-phase slug behavior, such
reduction in the transition from a slug or annular flow to stratified as liquid holdup, slug frequency, and length, along with wave
14
Fig. 11. A comparison of slug translational velocity vs. mixture velocity for several studies at different flow conditions (Baba et al., 2019).
behaviors at the gas–liquid interface. The viscosity effect has been 2014, Khaledi et al., 2014), and high translational velocity (Foletti
investigated experimentally by many researchers. (Dukler and et al., 2011, Losi and Poesio, 2016). However, the influence of the
Taitel, 1986) conducted experiments for gas–liquid two-phase flow liquid viscosity on the slug length and velocity is unclear and
in a horizontal pipe under various liquid viscosities of 1 cP, 90 cP, required ancillary investigation. Besides, most of the developed
and 165 cP. Reasonable agreement between their obtained results mechanistic models required hypothetical input parameters to
and the results of their previously developed model (Taitel and perform the prediction process (Losi et al., 2016).
Dukler, 1976) was achieved for liquid with a viscosity of 165 cP. The slug body flow is usually presumed fully developed, which
Also, (Andritsos et al., 1989) proposed a new slug initiation mech- implies a steady wall shear stress in the slug body. However,
anism, in which the slug formed and initiated after the presence of (Buitrago et al., 2017) found for high liquid viscosity, the wall shear
KH waves with a small wavelength. From their study, the slug sta- stress increased in the slug body and decrease in the liquid film.
bilized with increasing the liquid viscosities. Whereas (Gokcal Besides, they reported that a longer length is required to scoop
et al., 2006) compared their experimental results of slightly higher the liquid film upstream and induce an elongated mixing zone at
liquid viscosity (up to 587 cP) with models developed by (Xiao the slug front. Then, the slug body liquid mass dragged contained
et al., 1990, Zhang et al., 2003). They concluded that the accuracy liquid and continuously increased the wall shear stress. Mean-
of these models could be enhanced by considering liquid viscosity. while, the shear stresses decreased at the tail of the slug body
However, for the liquid with high viscosity, no validation studies due to flow deceleration. Despite these investigations, there is a
were carried out. Later, Gokcal (2008) found the Reynolds number lack of comprehensive flow field information, such as the velocity
decreased with increasing the viscosity, which might damp the field for highly viscous liquids. Although numerous studies have
turbulence of the mixing zone at the slug front. As a result, the slug been conducted to investigate the viscosity effect on slug flows,
length decreased with increasing liquid viscosity along with the however, the available experimental data and validated correla-
mixing zone length reduction. This conclusion was contradicted tions for high liquid viscosity are scarce (Zhao et al., 2015). Also,
by the finding of (Buitrago et al., 2017). Thus, further investigations the adopted methods through the previous mechanistic models
of slug length and the mixing zone under various liquid viscosities such as Dukler and Hubbard (1975), (Gomez, 1999), and Zhang
need to be ascertained. By contrast, (Marquez and Trujillo, 2010) et al. (2003 a, b) required additional modifications to be applied
tested the flow regime transition with liquid viscosity of 10 cP, for highly viscous slug flow (Pineda-Pérez et al., 2018). Recently,
392 cP, and 1000 cP. They discovered that the slug flow regime (Baba et al., 2019) developed an empirical correlation for slug
range on the map increased with increasing the liquid viscosity. translational velocity. The proposed correlation demonstrated
Moreover, a similar trend was observed by (Matsubara and immense prediction capabilities in predicting the slug velocity
Naito, 2011) for high liquid viscosity of 11,000 cP. They concluded for oil with a viscosity between 1 and 6 cP. They found that the
that roll wave and stratified flow at low liquid viscosity in the flow translational velocity increased with increasing the liquid viscos-
regime map were substituted by the intermittent and annular flow. ity, as shown in Fig. 11. However, they recommended expanding
Moreover, (Eskerud Smith et al., 2011) conducted a large-scale the data set and the operation conditions such as fluid properties
experiment on a pipe with 69 mm ID and 51.9 m length for high and pipe diameter to ensure the developed correlation’s
and low viscous oil. The oil’s viscosity within the range of 0.08– robustness.
0.11 Pa. s, which approximately 100 times denser than water. Their
results revealed that liquid fraction decreases significantly with
increased gas Reynolds number (ReL) at constant liquid Reynolds 6. Conclusion
number (ReL). Consequently, the liquid phase velocity increased
four times the highest gas flow rate. Therefore, the liquid passed This literature article is compiling the previous arts on gas–liq-
through the transition from laminar to turbulent liquid flow due uid slug flow in a horizontal pipe. It elucidates the developed the-
to the increase of the gas Reynolds number. Ancillary details about oretical models and empirical correlations that have been used to
the flow regime transition’s viscosity effect are found in (Barnea, determine the slug characteristics. Most of the previously devel-
1991, Andritsos and Hanratty, 1987a). Viscous slug flow was char- oped correlations are applicable to limited operational conditions,
acterized by the high frequency (Gokcal et al., 2009), short length pipe materials, and pipe diameters. They are generating results
(Al-Safran et al., 2011a, Brito, 2012, Kim et al., 2018), and a high with high discrepancies when utilized at different circumstances
liquid holdup (Al-Safran et al., 2013a), a high liquid holdup in slug from those at which developed. Therefore, a comprehensive model
film (Nädler and Mewes, 1995, Kora et al., 2011, Farsetti et al., that can be used for various operational conditions, fluid proper-
15
ties, pipe diameter, and material is highly desired. There is a scar- translational velocity in a horizontal pipe. Int. J. Heat Mass Transf. 105, 252–
260.
city of available experimental data and validated correlations for
Al-Lababidi, S., 2006. Multiphase Flow Measurement in the Slug Regime Using
high liquid, viscous slug flow. Besides, it is proven that there is a Ultrasonic Measurement Techniques and Slug Closure Model. Cranfield
lack of a comprehensive flow field and experimental laboratory University.
information for high liquid viscosity, in addition to the impact of Al-Safran, E., 2009. Investigation and prediction of slug frequency in gas/liquid
horizontal pipe flow. J. Petrol. Sci. Eng. 69, 143–155.
drag reduction on slug flow. By contrast, a full-scale experiment Al-Safran, E., Gokcal, B., Sarica, C., 2011. High viscosity liquid effect on two-phase
is required to investigate the bubble-centering phenomenon as it slug length in horizontal pipes. In: 15th International Conference on Multiphase
was hypothesized to have an immense role in the dynamic of slug Production Technology. BHR Group.
Al-Safran, E., Kora, C., Sarica, C., 2013. Prediction of liquid volume fraction in slugs in
flow with relatively high liquid viscosity. On the other hand, the two-phase horizontal pipe flow with high viscosity liquid. In: 16th International
slug flow is formed and initiated as a result of various mechanisms Conference on Multiphase Production Technology. BHR Group.
such as wave interaction and long-wave stability theory. Therefore, Al-Safran, E.M., 2016. Probabilistic modeling of slug frequency in gas/liquid pipe
flow using the Poisson probability theory. J. Petrol. Sci. Eng. 138, 88–96.
additional mathematical models are required to predict the transi- Al-Safran, E.M., Gokcal, B., Sarica, C., 2011. Analysis and prediction of heavy oil two-
tion lines between the flow regimes particularly, the transition phase slug length in horizontal pipelines. In: SPE Heavy Oil Conference and
from stratified flow to slug flow. Intelligence computational meth- Exhibition. Society of Petroleum Engineers.
Al-Safran, E.M., Gokcal, B., Sarica, C., 2013b. Investigation and prediction of high-
ods are used to identify the flow regime and slug characteristics viscosity liquid effect on two-phase slug length in horizontal pipelines. SPE
using Artificial Neural Networks (ANN). However, more attempts Prod. Oper. 28, 296–305.
are required to enhance prediction accuracy, computational cost, Al-Safran, E.M., Taitel, Y., Brill, J.P., 2004. Prediction of slug length distribution along
a hilly terrain pipeline using slug tracking model. J. Energy Resour. Technol. 126,
and computational slug analysis. The usage of these artificial intel-
54–62.
ligence methods could be a suitable route to eliminate the previous Al-Sarkhi, A., 2010. Drag reduction with polymers in gas-liquid/liquid-liquid flows
empirical correlations’ subjective utilization. in pipes: a literature review. J. Nat. Gas Sci. Eng. 2, 41–48.
It is recommended to enhance the understanding and predic- Al-Sarkhi, A., Hanratty, T., 2001. Effect of drag-reducing polymers on annular gas–
liquid flow in a horizontal pipe. Int. J. Multiph. Flow 27, 1151–1162.
tion of slug flow parameters in pipe flow by extra experimental, Al-Sarkhi, A., Soleimani, A., 2004. Effect of drag reducing polymers on two-phase
validated numerical, and empirical results. The authors suggest gas–liquid flows in a horizontal pipe. Chem. Eng. Res. Des. 82, 1583–1588.
in-depth scrutiny of the pressure and fluid–structure interaction Alsurakji, I.H., Al-Sarkhi, A., Habib, M., Badr, H.M., 2018. An experimental study on
the performance of drag-reducing polymers in single-and multiphase
of slug flow with the internal pipe wall, as there is a noticeable horizontal flow using particle image velocimetry. J. Energy Res. Technol. 140.
gap in this field. Those recommended studies are necessary to sup- Andreussi, P., Bendiksen, K., 1989. An investigation of void fraction in liquid slugs
port industrial designer’s and operators’ decisions in piping sys- for horizontal and inclined gas-liquid pipe flow. Int. J. Multiph. Flow 15, 937–
946.
tems’ design and operation. Moreover, additional experimental Andreussi, P., Pintus, S., Nydal O.J., Sanchez Silva, F., 1991. Measurement of the
investigations about the hydrodynamic behavior at the transition mean gas velocity in horizontal bubbly flow.
zone are recommended to develop supported correlations, which Andrianto, M., Widyaparaga, A., Dinaryanto, O., 2016. CFD studies on the gas-liquid
plug two-phase flow in a horizontal pipe. J. Petrol. Sci. Eng. 147, 779–787.
can ameliorate the understanding of the slug initiation, formation, Andritsos, N., Hanratty, T., 1987a. Influence of interfacial waves in stratified gas-
and transition. Immense endeavors are required to enhance under- liquid flows. AIChE J. 33, 444–454.
standing the local behavior of the two-phase drag-reducing flow Andritsos, N., Hanratty, T., 1987b. Interfacial instabilities for horizontal gas-liquid
flows in pipelines. Int. J. Multiph. Flow 13, 583–603.
induced by polymer additives.
Andritsos, N., Williams, L., Hanratty, T., 1989. Effect of liquid viscosity on the
Further investigations are recommended to resolve the discrep- stratified-slug transition in horizontal pipe flow. Int. J. Multiph. Flow 15, 877–
ancies and inherent uncertainty of the developed slug frequency 892.
correlations. These investigations should consider the abrupt shift Ansari, M., Shokri, V., 2011. Numerical modeling of slug flow initiation in a
horizontal channels using a two-fluid model. Int. J. Heat Fluid Flow 32, 145–
between the slug formation mechanisms within the same opera- 155.
tional parameters. On the contrary, the existent numerical models Antunes, M.M., Cozin, C., Barbuto, F.A., Morales, R.E., Rodrigues, H.T., 2014. Analysis
can predict the slug characteristic accurately; however, the compu- of slug frequency correlations for two-phase gas-liquid horizontal slug flow. In:
ASME 2014 4th Joint US-European Fluids Engineering Division Summer
tational cost is yet high. Therefore, a robust numerical model is Meeting collocated with the ASME 2014 12th International Conference on
required to simulate the slug flow precisely. This model must pre- Nanochannels, Microchannels, and Minichannels. American Society of
dict the general and the local characteristic behaviors of the slug Mechanical Engineers Digital Collection.
Arabi, A., Salhi, Y., Zenati, Y., Si-Ahmed, E., Legrand, J., 2020. On gas-liquid
flow accurately with less computational time. intermittent flow in a horizontal pipe: Influence of sub-regime on slug
frequency. Chem. Eng. Sci. 211, 115251.
Azizi, H.Z., Naghashzadegan, M., Shokri, V., 2019. Comparison of numerical methods
Declaration of Competing Interest for two-fluid model for gas-liquid transient flow regime and its application in
slug modeling initiation. Iran. J. Sci. Technol., Trans. Mech. Eng. 43, 663–673.
The authors declare that they have no known competing finan- Baba, Y.D., Aliyu, A.M., Archibong, A.E., Abdulkadir, M., Lao, L., Yeung, H., 2018. Slug
length for high viscosity oil-gas flow in horizontal pipes: Experiments and
cial interests or personal relationships that could have appeared prediction. J. Petrol. Sci. Eng. 165, 397–411.
to influence the work reported in this paper. Baba, Y.D., Archibong-Eso, A., Aliyu, A.M., Fajemidupe, O.T., Ribeiro, J.X., Lao, L.,
Yeung, H., 2019. Slug translational velocity for highly viscous oil and gas flows
in horizontal pipes. Fluids 4, 170.
References Barmak, I., Gelfgat, A., Vitoshkin, H., Ullmann, A., Brauner, N., 2016. Stability of
stratified two-phase flows in horizontal channels. Phys. Fluids 28, 044101.
Barnea, D., 1991. On the effect of viscosity on stability of stratified gas-liquid flow-
Abdulkadir, M., Hernandez-Perez, V., Lowndes, I., Azzopardi, B., Sam-Mbomah, E.,
application to flow pattern transition at various pipe inclinations. Chem. Eng.
2016. Experimental study of the hydrodynamic behaviour of slug flow in a
Sci. 46, 2123–2131.
horizontal pipe. Chem. Eng. Sci. 156, 147–161.
Barnea, D., Brauner, N., 1985. Holdup of the liquid slug in two phase intermittent
Akbar, M., Ghiaasiaan, S., 2006. Simulation of Taylor flow in capillaries based on the
flow. Int. J. Multiph. Flow 11, 43–49.
volume-of-fluid technique. Ind. Eng. Chem. Res. 45, 5396–5403.
Barnea, D., Taitel, Y., 1993a. Kelvin-Helmholtz stability criteria for stratified flow:
Akhlaghi, M., Mohammadi, V., Nouri, N.M., Taherkhani, M., Karimi, M., 2019. Multi-
viscous versus non-viscous (inviscid) approaches. Int. J. Multiph. Flow 19, 639–
Fluid VoF model assessment to simulate the horizontal air–water intermittent
649.
flow. Chem. Eng. Res. Des. 152, 48–59.
Barnea, D., Taitel, Y., 1993b. A model for slug length distribution in gas-liquid slug
Al-Hashimy, Z.I., Al-Kayiem, H.H., Nasif, M.S., Mohmmed, A., 2016. Numerical
simulation of the transient development of slug flow in horizontal pipes. Appl.
Bendiksen, K., Espedal, M., 1992. Onset of slugging in horizontal gas-liquid pipe
Mech. Mater., Trans Tech Publ, pp. 300–304.
Al-Kayiem, H.H., Al-Hashimy, Z.I., Time, R.W., Kadhim, Z.K. Comparative
Bendiksen, K.H., 1984. An experimental investigation of the motion of long bubbles
experimental and numerical investigations on the characteristics of air/water
in inclined tubes. Int. J. Multiph. Flow 10, 467–483.
slug flow in horizontal pipes.
Bendiksen, K.H., 1985. On the motion of long bubbles in vertical tubes. Int. J.
Al-Kayiem, H.H., Mohmmed, A.O., Al-Hashimy, Z.I., Time, R.W., 2017. Statistical
Multiph. Flow 11, 797–812.
assessment of experimental observation on the slug body length and slug
16
Bendiksen, K.H., Langsholt, M., Liu, L., 2018. An experimental investigation of the Dukler, A.E., Maron, D.M., Brauner, N., 1985. A physical model for predicting the
motion of long bubbles in high viscosity slug flow in horizontal pipes. Int. J. minimum stable slug length. Chem. Eng. Sci. 40, 1379–1385.
Multiph. Flow 104, 60–73. Dumitrescu, D.T., 1943. Strömung an einer Luftblase im senkrechten Rohr. ZAMM-J.
Benjamin, T.B., 1968. Gravity currents and related phenomena. J. Fluid Mech. 31, Appl. Math. Mech./Zeitschrift für Angewandte Mathematik und Mechanik 23,
209–248. 139–149.
Bernicot, M., Drouffe, J.-M., 1991. A slug-length distribution law for multiphase Eskerud Smith, I., Krampa, F., Fossen, M., Brekken, C., Unander, T., 2011.
transportation systems. SPE Prod. Eng. 6, 166–170. Investigation of horizontal two-phase gas-liquid pipe flow using high
Bernicot, M., Drouffe, J., 1989. Slug length distribution in diphasic transportation viscosity oil: comparison with experiments using low viscosity oil and
systems. In: Proceedings of the 4th International Conference on Multiphase simulations. BHR Group.
Flow, Nice. pp. 19–21. Eyo, E.N., Lao, L., 2019. Slug flow characterization in horizontal annulus. AIChE J. 65,
Bertola, V., 2002. Slug velocity profiles in horizontal gas–liquid flow. Exp. Fluids 32, e16711.
722–727. Fabre, J., 2003. Gas-liquid slug flow. Modelling and experimentation in two-phase
Bertola, V., 2003. Experimental characterization of gas–liquid intermittent flow. Springer.
subregimes by phase density function measurements. Exp. Fluids 34, 122–129. Fabre, J., Liné, A., 1992. Modeling of two-phase slug flow. Annu. Rev. Fluid Mech. 24,
Bonizzi, M., 2003. Transient one-dimensional modelling of multiphase slug flows. 21–46.
Bonizzi, M., Issa, R., 2003. On the simulation of three-phase slug flow in nearly Fagundes Netto, J., Fabre, J., Grenier, P., Péresson, L., 1998. An experimental study of
horizontal pipes using the multi-fluid model. Int. J. Multiph. Flow 29, 1719– an isolated long bubble in an horizontal liquid flow. In: Third International
1747. Conference on Multiphase Flow, ICMF, pp. 8–12.
Bontozoglou, V., Hanratty, T., 1989. Wave height estimation in stratified gas-liquid Fan, Z., Lusseyran, F., Hanratty, T., 1993. Initiation of slugs in horizontal gas-liquid
flows. AIChE J. 35, 1346–1350. flows. AIChE J. 39, 1741–1753.
Brill, J.P.Schmidt, Z.R., Coberly, W.A., Herring, J.D. and Moore, D.W., 1981. Analysis of Farsetti, S., Farisè, S., Poesio, P., 2014. Experimental investigation of high viscosity
two-phase tests in large diameter flow lines in Prudhoe Bay field. Sot. Pet. Eng. J oil–air intermittent flow. Exp. Therm Fluid Sci. 57, 285–292.
3, 363–378. https://doi.org/10.2118/8305-PA. In this issue. Ferrari, M., Bonzanini, A., Poesio, P., 2019. A slug capturing method in
Brito, R., 2012. Effect of Medium Oil Viscosity On Two-Phase Oil-Gas Flow Behavior unconventional scenarios: The 5ESCARGOTS code applied to non-Newtonian
In Horizontal Pipes. University of Tulsa. fluids, high viscous oils and complex geometries. Petroleum 5, 171–177.
Buitrago, L., Kim, T., Pereyra, E., Sarica, C., 2017. Wall shear stress measurements of Ferre, D., 1978. Experimental study of two-phase flows with pockets in horizontal
horizontal two-phase slug flow for high viscosity liquids using constant pipes.
temperature anemometry. In: 18th International Conference on Multiphase Fetter, C., 1988. Development of a clamp-on acoustic two-phase flowmeter.
Production Technology. BHR Group. Foletti, C., Farisè, S., Grassi, B., Strazza, D., Lancini, M., Poesio, P., 2011. Experimental
Carlson, A., Kudinov, P., Narayanan, C., 2008. Prediction of two-phase flow in small investigation on two-phase air/high-viscosity-oil flow in a horizontal pipe.
tubes: a systematic comparison of state-of-the-art cmfd codes. In: 5th European Chem. Eng. Sci. 66, 5968–5975.
Thermal-Sciences Conference, The Netherlands, pp. 126–130. Fossa, M., Guglielmini, G., Marchitto, A., 2003. Intermittent flow parameters from
Carneiro, J., Fonseca Jr, R., Ortega, A., Chucuya, R., Nieckele, A., Azevedo, L., 2011. void fraction analysis. Flow Meas. Instrum. 14, 161–168.
Statistical characterization of two-phase slug flow in a horizontal pipe. J. Braz. Frank, T., 2005. Numerical simulation of slug flow regime for an air-water two-
Soc. Mech. Sci. Eng. 33, 251–258. phase flow in horizontal pipes. In: Proceedings of the 11th International Topical
Cerne, G., Petelin, S., Tiselj, I., 2001. Coupling of the interface tracking and the two- Meeting on Nuclear Reactor Thermal-Hydraulics (NURETH-11), Avignon,
fluid models for the simulation of incompressible two-phase flow. J. Comput. France, October, 2005. pp. 2–6.
Phys. 171, 776–804. Fukagata, K., Kasagi, N., Ua-Arayaporn, P., Himeno, T., 2007. Numerical simulation of
Choi, J., Pereyra, E., Sarica, C., Park, C., Kang, J.M., 2012. An efficient drift-flux closure gas–liquid two-phase flow and convective heat transfer in a micro tube. Int. J.
relationship to estimate liquid holdups of gas-liquid two-phase flow in pipes. Heat Fluid Flow 28, 72–82.
Energies 5, 5294–5306. Ghorai, S., Nigam, K., 2006. CFD modeling of flow profiles and interfacial
Collins, R., de Moraes, F., Davidson, J., Harrison, D., 1978. The motion of a large gas phenomena in two-phase flow in pipes. Chem. Eng. Process. Process Intensif.
bubble rising through liquid flowing in a tube. J. Fluid Mech. 89, 497–514. 45, 55–65.
Conte, M.G., Hegde, G.A., da Silva, M.J., Sum, A.K., Morales, R.E., 2017. Godino, D.M., Corzo, S.F., Ramajo, D.E., 2020. Two-phase modeling of water-air flow
Characterization of slug initiation for horizontal air-water two-phase flow. of dispersed and segregated flows. Ann. Nucl. Energy 149, 107766.
Exp. Therm Fluid Sci. 87, 80–92. Gokcal, B., Al-Sarkhi, A., Sarica, C., Al-Safran, E.M., 2010. Prediction of slug frequency
Crowley, C., Wallis, G., Barry, J., 1992. Validation of a one-dimensional wave model for high-viscosity oils in horizontal pipes. SPE Projects, Facil. Constr. 5, 136–144.
for the stratified-to-slug flow regime transition, with consequences for wave Gokcal, B., Al-Sarkhi, A.S., Sarica, C., 2009. Effects of high oil viscosity on drift velocity
growth and slug frequency. Int. J. Multiph. Flow 18, 249–271. for horizontal and upward inclined pipes. SPE Projects, Facil. Constr. 4, 32–40.
Daas, M., Bleyle, D., 2006. Computational and experimental investigation of the drag Gokcal, B., Wang, Q., Zhang, H., Sarica, C., 2006. Effects of high oil viscosity on oil/gas
reduction and the components of pressure drop in horizontal slug flow using flow behavior in horizontal pipes. SPE Annu. Tech. Conf. Exhib, 24–27.
liquids of different viscosities. Exp. Therm Fluid Sci. 30, 307–317. Gomez, L.E.O. Shoham Z. Schmidt R.N. Chokshi A. Brown T. Northug, 1999. A Unified
Daas, M., Kang, C., Jepson, W., 2000. Quantitative analysis of drag reduction in Mechanistic Model for Steady-State Two-Phase Flow in Wellbores and
horizontal slug flow. In: SPE Annual Technical Conference and Exhibition. Pipelines. SPE Annual Technical Conference and Exhibition. https://doi.org/
Society of Petroleum Engineers. 10.2118/56520-MS.
Danielson, T.J., Bansal, K.M., Djoric, B., Larrey, D., Johansen, S.T., De Leebeeck, A., Gong, Z., 2010. Explosion analysis of the steam pipeline. Guangzhou Chem. Ind. 38,
Kjolaas, J., 2012. Simulation of slug flow in oil and gas pipelines using a new 181–182.
transient simulator. In: Offshore Technology Conference, 2012. Offshore Gordon, I., Fairhurst, C., 1988. Multi-phase pipeline and equipment design for
Technology Conference. marginal and deep-water field development. Pipes Pipelines Int. 1965 (33), 13–21.
Davies, R., Taylor, G.I., 1950. The mechanics of large bubbles rising through Gregory, G., Nicholson, M., Aziz, K., 1978. Correlation of the liquid volume fraction
extended liquids and through liquids in tubes. Proc. R. Soc. Lond. A 200, 375– in the slug for horizontal gas-liquid slug flow. Int. J. Multiph. Flow 4, 33–39.
390. Gregory, G., Scott, D., 1969. Correlation of liquid slug velocity and frequency in
de Henau, V., Raithby, G., 1995a. A transient two-fluid model for the simulation of horizontal cocurrent gas-liquid slug flow. AIChE J. 15, 933–935.
slug flow in pipelines-I. Theory. Int. J. Multiphase Flow 21, 335–349. Grenier, P., Fabre, J., Fagundes Netto, J., 1997. Slug flow in pipelines: recent advances
de Henau, V., Raithby, G., 1995b. A transient two-fluid model for the simulation of and future developments. BHR Group Conference Series Publication, 1997.
slug flow in pipelines-II. Validation. Int. J. Multiphase Flow 21, 351–363. Mechanical Engineering Publications Limited, pp. 107–124.
de Leebeeck, A., Nydal, O., 2010. Simulation of large amplitude waves in a slug Greskovich, E., Shrier, A., 1971. Pressure drop and holdup in horizontal slug flow.
tracking scheme compared to roll wave experiments at high pressure. Int. J. AIChE J. 17, 1214–1219.
Multiph. Flow 36, 40–50. Guerrero, E., Muñoz, F., Ratkovich, N., 2017. Comparison between Eulerian and VOF
Denèfle, R., Mimouni, S., Caltagirone, J.-P., Vincent, S., 2015. Multifield hybrid models for two-phase flow assessment in vertical pipes. CT&F-Ciencia,
approach for two-phase flow modeling–Part 1: adiabatic flows. Comput. Fluids Tecnología y Futuro 7, 73–84.
113, 106–111. Gupta, R., Fletcher, D.F., Haynes, B.S., 2009. On the CFD modelling of Taylor flow in
Dhulesia, H., Bernicot, M., Deheuvels, P., 1991. Statistical analysis and modelling of microchannels. Chem. Eng. Sci. 64, 2941–2950.
slug lengths. In: Proceedings of the ÿfth international conference on multiphase Hale, C.P., 2001. Slug formation, growth and decay in gas-liquid flows.
production, Cannes, France, BHRA, Cranfield, Beds. pp. 80–112. Han, G.-Q., Shi, S.-Z., Wu, X.-D., You, C.-C., Cen, X.-Q., Zhang, Z.-G., Zhu, Z.-Y., Sun, K.,
Dhulesia, H., Hustvedt, E., Todal, O., 1993. Measurement and analysis of slug Tao, Y.-C., 2019. Characteristics analysis and slug flow model in undulated
characteristics in multiphase pipelines. Proc. wellbore. In: International Petroleum and Petrochemical Technology
Dinaryanto, O., Prayitno, Y.A.K., Majid, A.I., Hudaya, A.Z., Nusirwan, Y.A., Conference. Springer, pp. 190–203.
Widyaparaga, A., 2017. Experimental investigation on the initiation and flow He, Q., Kasagi, N., 2008. Phase-field simulation of small capillary-number two-phase
development of gas-liquid slug two-phase flow in a horizontal pipe. Exp. Therm flow in a microtube. Fluid Dyn. Res. 40, 497.
Fluid Sci. 81, 93–108. Heywood, N., Richardson, J., 1979. Slug flow of air-water mixtures in a horizontal
Dukler, A., Taitel, Y., 1986. Flow pattern transitions in gas-liquid systems: pipe: Determination of liquid holdup by c-ray absorption. Chem. Eng. Sci. 34,
measurement and modeling. Multiphase Sci. Technol. 2. 17–30.
Dukler, A.E., Fabre, J., 1994. Gas-liquid slug flow. Multiphase. Sci. Technol. 8. Hill, T., Fairhurst, C., Nelson, C., Becerra, H., Bailey, R., 1996. Multiphase production
Dukler, A.E., Hubbard, M.G., 1975. A model for gas-liquid slug flow in horizontal and through hilly-terrain pipelines in Cusiana oilfield, Colombia, SPE 36606. SPE
near horizontal tubes. Ind. Eng. Chem. Fundam. 14, 337–347. ATCE, Denver, CO.
17
Hill, T., Wood, D., 1990. A new approach to the prediction of slug frequency. In: SPE Liu, L., 2013. A new method for evaluating drag reduction in gas–liquid two-phase
Annual Technical Conference and Exhibition. Society of Petroleum Engineers. flow based on energy dissipation. Chem. Eng. Sci. 95, 54–64.
Hill, T., Wood, D., 1994. Slug flow: Occurrence, consequences, and prediction. In: López, J., Pineda, H., Bello, D., Ratkovich, N., 2016. Study of liquid–gas two-phase
University of Tulsa Centennial Petroleum Engineering Symposium. Society of flow in horizontal pipes using high speed filming and computational fluid
Petroleum Engineers. dynamics. Exp. Therm Fluid Sci. 76, 126–134.
Hirt, C.W., Amsden, A.A., Cook, J., 1974. An arbitrary Lagrangian-Eulerian computing Losi, G., Arnone, D., Correra, S., Poesio, P., 2016. Modelling and statistical analysis of
method for all flow speeds. J. Comput. Phys. 14, 227–253. high viscosity oil/air slug flow characteristics in a small diameter horizontal
Hirt, C.W., Nichols, B.D., 1981. Volume of fluid (VOF) method for the dynamics of pipe. Chem. Eng. Sci. 148, 190–202.
free boundaries. J. Comput. Phys. 39, 201–225. Losi, G., Poesio, P., 2016. An experimental investigation on the effect of viscosity on
Horgue, P., Augier, F., Quintard, M., Prat, M., 2012. A suitable parametrization to bubbles moving in horizontal and slightly inclined pipes. Exp. Therm Fluid Sci.
simulate slug flows with the Volume-Of-Fluid method. Comptes Rendus 75, 77–88.
Mécanique 340, 411–419. Lu, G.-Y., Wang, J., Jia, Z.-H., 2007. Experimental and numerical investigations on
Hubbard, M.G., 1965. An analysis of horizontal gas-liquid slug flow, Tex.] University horizontal oil-gas flow. J. Hydrodyn. 19, 683–689.
of Houston. Lu, M., 2015. Experimental and Computational Study of Two-Phase Slug Flow.
Hurlburt, E., Hanratty, T., 2002. Prediction of the transition from stratified to slug Imperial College London.
and plug flow for long pipes. Int. J. Multiph. Flow 28, 707–729. Manfield, P.D., 2000. Experimental, computational and analytical studies of slug
Ishii, M., Mishima, K., 1984. Two-fluid model and hydrodynamic constitutive flow.
relations. Nucl. Eng. Des. 82, 107–126. Manolis, I., Mendes-Tatsis, M., Hewitt, G., 1995. The effect of pressure on slug
Issa, R., Kempf, M., 2003. Simulation of slug flow in horizontal and nearly horizontal frequency in two-phase horizontal flow. In: Multiphase Flow 1995. Elsevier.
pipes with the two-fluid model. Int. J. Multiph. Flow 29, 69–95. Manolis, I.G., 1995. High Pressure Gas-Liquid Slug Flow. Imperial College London.
Jepson, W., Taylor, R., 1993. Slug flow and its transitions in large-diameter Marquez, J.G., Trujillo, J.N., 2010. Overview: slug-flow characterization for heavy-oil
horizontal pipes. Int. J. Multiph. Flow 19, 411–420. fields. In: SPE Latin American and Caribbean Petroleum Engineering Conference.
Johansen, S.T., 2019. Method for simulation of multiphase fluid flow in pipelines. Society of Petroleum Engineers.
Google Patents. Matsubara, H., Naito, K., 2011. Effect of liquid viscosity on flow patterns of gas-
Kadri, U., Mudde, R., Oliemans, R., 2010. Influence of the operation pressure on slug liquid two-phase flow in a horizontal pipe. Int. J. Multiph. Flow 37, 1277–1281.
length in near horizontal gas–liquid pipe flow. Int. J. Multiph. Flow 36, 423–431. Mishima, K., Ishii, M. 1980. Theoretical prediction of onset of horizontal slug flow.
Kadri, U., Mudde, R., Oliemans, R., Bonizzi, M., Andreussi, P., 2009a. Prediction of the Mo, S., Ashrafian, A., Barbier, J.-C., Johansen, S.T., 2014. Quasi-3D modelling of two-
transition from stratified to slug flow or roll-waves in gas–liquid horizontal phase slug flow in pipes. J. Comput. Multiphase Flows 6, 1–12.
pipes. Int. J. Multiph. Flow 35, 1001–1010. Mohmmed, A., 2016. Effect of Slug Two-Phase Flow on Fatigue of Pipe Material.
Kadri, U., Zoeteweij, M., Mudde, R., Oliemans, R., 2009b. A growth model for Department of Mechanical Engineering, Universiti Teknologi, Petronas.
dynamic slugs in gas–liquid horizontal pipes. Int. J. Multiph. Flow 35, 439–449. Mohmmed, A.O., Al-Kayiem, H.H., Nasif, M.S., Time, R.W., 2019. Effect of slug flow
Kang, C., Jepson, W., 1999. Multiphase flow conditioning using drag-reducing frequency on the mechanical stress behavior of pipelines. Int. J. Press. Vessels
agents. In: SPE Annual Technical Conference and Exhibition. Society of Pip. 172, 1–9.
Petroleum Engineers. Mohmmed, A.O., Al-Kayiem, H.H., Osman, A., Sabir, O., 2020. One-way coupled
Kang, C., Jepson, W., 2000. Effect of drag-reducing agents in multiphase, oil/gas fluid–structure interaction of gas–liquid slug flow in a horizontal pipe:
horizontal flow. In: SPE International Petroleum Conference and Exhibition in experiments and simulations. J. Fluids Struct. 97, 103083.
Mexico. Society of Petroleum Engineers. Mohmmed, A.O., Nasif, M.S., Al-Kayiem, H.H., 2018. Numerical investigation of slug
Kang, C., Jepson, W., Gopal, M., 1999. Effect of drag-reducing agent on slug characteristics in a horizontal air/water and air/oil pipe flow. Prog. Comput.
characteristics in multiphase flow in inclined pipes. Fluid Dyn., Int. J. 18, 241–256.
Kashid, M.N., Rivas, D.F., Agar, D., Turek, S., 2008. On the hydrodynamics of liquid– Montini, M., 2011. Closure relations of the one-dimensional two-fluid model for the
liquid slug flow capillary microreactors. Asia-Pac. J. Chem. Eng. 3, 151–160. simulation of slug flows.
Khaledi, H.A., Smith, I.E., Unander, T.E., Nossen, J., 2014. Investigation of two-phase Mowla, D., Naderi, A., 2006. Experimental study of drag reduction by a polymeric
flow pattern, liquid holdup and pressure drop in viscous oil–gas flow. Int. J. additive in slug two-phase flow of crude oil and air in horizontal pipes. Chem.
Multiph. Flow 67, 37–51. Eng. Sci. 61, 1549–1554.
Kim, T.-W., Woo, N.-S., Han, S.-M., Kim, Y.-J., 2020. Optimization and extended Nädler, M., Mewes, D., 1995. Effects of the liquid viscosity on the phase distributions
applicability of simplified slug flow model for liquid-gas flow in horizontal and in horizontal gas-liquid slug flow. Int. J. Multiph. Flow 21, 253–266.
near horizontal pipes. Energies 13, 842. Naidek, B.P., Conte, M.G., Cozin, C., Barbuto, F.A., Morales, R.E., 2017. Performance
Kim, T., Aydin, T., Pereyra, E., Sarica, C., 2018. Detailed flow field measurements and analysis of the slug tracking modeling for intermittent flows in horizontal pipes
analysis in highly viscous slug flow in horizontal pipes. Int. J. Multiph. Flow 106, with long lengths. In: ASME 2017 International Mechanical Engineering
75–94. Congress and Exposition. American Society of Mechanical Engineers Digital
King, M.J.S., 1998. Experimental and Modelling Studies of Transient Slug Flow. Collection.
Imperial College London (University of London). Nesyn, G.V., Sunagatullin, R., Shibaev, V., Malkin, A.Y., 2018. Drag reduction in
Kjølaas, J., De Leebeeck, A., Johansen, S., 2013. Simulation of hydrodynamic slug flow transportation of hydrocarbon liquids: From fundamentals to engineering
using the LedaFlow slug capturing model. In: 16th International Conference on applications. J. Petrol. Sci. Eng. 161, 715–725.
Multiphase Production Technology. BHR Group. Nicholson, M., Aziz, K., Gregory, G., 1978. Intermittent two phase flow in horizontal
Kjølaas, J., Unander, T.E., Wolden, M., Schümann, H., Leinan, P.R., Smith, I.E., pipes: predictive models. Canadian J. Chem. Eng. 56, 653–663.
Shmueli, A., 2020. Large scale experiments on slug length evolution in long Nicklin, D., 1962. Two-phase bubble flow. Chem. Eng. Sci. 17, 693–702.
pipes. In: Offshore Technology Conference, 2020. Offshore Technology Nieckele, A., Carneiro, J., Chucuya, R., Azevedo, J., 2013. Initiation and statistical
Conference. evolution of horizontal slug flow with a two-fluid model. J. Fluids Eng. 135.
Kong, R., Rau, A., Kim, S., Bajorek, S., Tien, K., Hoxie, C., 2018. Experimental study of Nydal, O., Banerjee, S., 1996. Dynamic slug tracking simulations for gas-liquid flow
horizontal air-water plug-to-slug transition flow in different pipe sizes. Int. J. in pipelines. Chem. Eng. Commun. 141, 13–39.
Heat Mass Transf. 123, 1005–1020. Nydal, O., Pintus, S., Andreussi, P., 1992. Statistical characterization of slug flow in
Kora, C., Sarica, C., Zhang, H.-Q., Al-Sarkhi, A., Al-Safran, E., 2011. Effects of high oil horizontal pipes. Int. J. Multiph. Flow 18, 439–453.
viscosity on slug liquid holdup in horizontal pipes. In: Canadian Unconventional Oliver, D., Hoon, A.Y., 1968. Two-phase non-newtonian flow. Part II: Heat transfer.
Resources Conference. Society of Petroleum Engineers. Trans. Instn Chem. Engrs 46, T116–T122.
Kordyban, E.S., Ranov, T., 1970. Mechanism of slug formation in horizontal two- Parsi, M., Agrawal, M., Srinivasan, V., Vieira, R.E., Torres, C.F., McLaury, B.S., Shirazi,
phase flow. S.A., Schleicher, E., Hampel, U., 2016. Assessment of a hybrid CFD model for
Kouba, G., Jepson, W., 1990. The flow of slugs in horizontal, two-phase pipelines. simulation of complex vertical upward gas–liquid churn flow. Chem. Eng. Res.
Lakehal, D., Labois, M., Narayanan, C., 2012. Advances in the large–eddy and Des. 105, 71–84.
interface simulation (leis) of interfacial multiphase flows in pipes. Prog. Petalas, N., Aziz, K., 2000. A mechanistic model for multiphase flow in pipes. J. Can.
Comput. Fluid Dyn., Int. J. 12, 153–163. Pet. Technol. 39.
Lakehal, D., Meier, M., Fulgosi, M., 2002. Interface tracking towards the direct Picchi, D., Manerba, Y., Correra, S., Margarone, M., Poesio, P., 2015. Gas/shear-
simulation of heat and mass transfer in multiphase flows. Int. J. Heat Fluid Flow thinning liquid flows through pipes: modeling and experiments. Int. J. Multiph.
23, 242–257. Flow 73, 217–226.
Laux, H., Meese, E., Johansen, S., Ladam, Y., Bansal, K., Danielson, T., Goldszal, A., Pineda-Pérez, H., Kim, T., Pereyra, E., Ratkovich, N., 2018. CFD modeling of air and
Monsen, J., 2007. Simulation of multiphase flows composed of large scale highly viscous liquid two-phase slug flow in horizontal pipes. Chem. Eng. Res.
interfaces and dispersed fields. Proceedings of the 6th International Conference Des. 136, 638–653.
on Multiphase Flow. Polonsky, S., Shemer, L., Barnea, D., 1999. The relation between the Taylor bubble
Lawrence, C.J., Hu, B., 2020. Slug flow initiation in fluid flow models. Google Patents. motion and the velocity field ahead of it. Int. J. Multiph. Flow 25, 957–975.
Li, F., An, C., Duan, M., Su, J., 2020. Combined damping model for dynamics and Pouranfard, A., Mowla, D., Esmaeilzadeh, F., 2015. An experimental study of drag
stability of a pipe conveying two-phase flow. Ocean Eng. 195, 106683. reduction by nanofluids in slug two-phase flow of air and water through
Lin, M., Liu, Y., Hu, Y., Che, D., 2013. Influence of the gas and liquid superficial horizontal pipes. Chin. J. Chem. Eng. 23, 471–475.
velocity on slug frequency. In: AIP Conference Proceedings. American Institute Qian, D., Lawal, A., 2006. Numerical study on gas and liquid slugs for Taylor flow in a
of Physics, pp. 253–263. T-junction microchannel. Chem. Eng. Sci. 61, 7609–7625.
Lin, P., Hanratty, T., 1986. Prediction of the initiation of slugs with linear stability Rahmandhika, A., Dinaryanto, O., Widyatama, A., Hudaya, A.Z., Indarto
theory. Int. J. Multiph. Flow 12, 79–98. Deendarlianto, 2018. Visualization study in the transition flow pattern of
18
stratified to slug flow of air-water two phase flow in a horizontal pipe. In: AIP Ujang, P.M., Lawrence, C.J., Hale, C.P., Hewitt, G.F., 2006. Slug initiation and
Conference Proceedings. AIP Publishing LLC, 030003. evolution in two-phase horizontal flow. Int. J. Multiph. Flow 32, 527–552.
Rahmandhika, A., Widyatama, A., Dinaryanto, O., Widyaparaga, A., 2019. van Hout, R., Barnea, D., Shemer, L., 2002. Translational velocities of elongated
Experimental study on the hydrodynamic behavior of gas-liquid air-water bubbles in continuous slug flow. Int. J. Multiph. Flow 28, 1333–1350.
two-phase flow near the transition to slug flow in horizontal pipes. Int. J. Heat Vaze, M., Banerjee, J., 2011. Experimental visualization of two-phase flow patterns
Mass Transf. 130, 187–203. and transition from stratified to slug flow. Proc. Inst. Mech. Eng., Part C: J. Mech.
Renault, F., 2007. A Lagrangian slug capturing scheme for gas-liquid flows in pipes. Eng. Sci. 225, 382–389.
Rodrigues, H.T., Morales, R.E., Mazza, R.A., Rosa, E., 2008. Horizontal slug flow Wallis, G., 1969. One-Dimensional Two-Phase Flow. McGraw-Hil Co., New York
properties captured by intermittent slug tracking model. In: 12th Brazilian City.
Congress of Thermal Engineering and Sciences, pp. 10–14. Wallis, G.B., Dodson, J.E., 1973. The onset of slugging in horizontal stratified air-
Rosa, E., Mazza, R., Morales, R., Rodrigues, H., Cozin, C., 2015. Analysis of slug water flow. Int. J. Multiph. Flow 1, 173–193.
tracking model for gas–liquid flows in a pipe. J. Braz. Soc. Mech. Sci. Eng. 37, Wan, T.T., Loh, W.L., Nguyen, D.T., Perez, V.H., Premanadhan, V.K., 2019. Diameter
1665–1686. scaling correlations for near atmospheric air-water slug velocities in horizontal
Ruder, Z., Hanratty, P., Hanratty, T., 1989. Necessary conditions for the existence of pipe. J. Petrol. Sci. Eng. 172, 349–359.
stable slugs. Int. J. Multiph. Flow 15, 209–226. Wang, L., Yang, Y., Li, Y., Wang, Y., 2018. Dynamic behaviours of horizontal gas-
Sanchis, A., Johnson, G.W., Jensen, A., 2011. The formation of hydrodynamic slugs by liquid pipes subjected to hydrodynamic slug flow: modelling and experiments.
the interaction of waves in gas–liquid two-phase pipe flow. Int. J. Multiph. Flow Int. J. Press. Vessels Pip. 161, 50–57.
37, 358–368. Wang, S.-Q., Xu, K.-W., Kim, H.-B., 2020. Slug flow identification using ultrasound
Schmelter, S., Knotek, S., Olbrich, M., Fiebach, A., Bär, M., 2021. On the influence of Doppler velocimetry. Int. J. Heat Mass Transf. 148, 119004.
inlet perturbations on slug dynamics in horizontal multiphase flow-a Wang, X., Guo, L., Zhang, X., 2007. An experimental study of the statistical
computational study. Metrologia 58, 014003. parameters of gas–liquid two-phase slug flow in horizontal pipeline. Int. J. Heat
Schulkes, R., 2011. Slug frequencies revisited. In: 15th International Conference on Mass Transf. 50, 2439–2443.
Multiphase Production Technology. BHR Group. Wang, Y., Wang, Y., Cheng, Z., 2019. Direct numerical simulation of gas-liquid drag-
Scott, S., Kouba, G., 1990. Advances in slug flow characterization for horizontal and reducing cavity flow by the VOSET method. Polymers 11, 596.
slightly inclined pipelines. SPE Annual Technical Conference and Exhibition. Wilkens, R., Thomas, D., 2008. A simple technique for determining slug frequency
Society of Petroleum Engineers. using differential pressure. J. Energy Res. Technol. 130.
Scott, Stuart LOvadia Shoham, James P Brill, 1989. Prediction of slug length in Woods, B., Hanratty, T., 1996. Relation of slug stability to shedding rate. Int. J.
horizontal, large-diameter pipes. SPE Production Engineering (03), 335–340. Multiph. Flow 22, 809–828.
https://doi.org/10.2118/15103-PA. Woods, B.D., Fan, Z., Hanratty, T.J., 2006. Frequency and development of slugs in a
Sharma, S., Lewis, S., Kojasoy, G., 1998. Local studies in horizontal gas–liquid slug horizontal pipe at large liquid flows. Int. J. Multiph. Flow 32, 902–925.
flow. Nucl. Eng. Des. 184, 305–318. Woods, B.D., Hanratty, T.J., 1999. Influence of Froude number on physical processes
Shea, R., Eidsmoen, H., Nordsveen, M., Rasmussen, J., Xu, Z., Nossen, J., 2004. Slug determining frequency of slugging in horizontal gas–liquid flows. Int. J.
frequency prediction method comparison. In: Proceedings of the 4th North Multiph. Flow 25, 1195–1223.
American Conference on Multiphase Technology, pp. 227–237. Woods, B.D., Hurlburt, E.T., Hanratty, T.J., 2000. Mechanism of slug formation in
Shemer, L., 1987. Visualization of the instantaneous velocity profiles in gas-liquid downwardly inclined pipes. Int. J. Multiph. Flow 26, 977–998.
slug flow. PhysicoChem. Hydrodyn. 8, 243–253. Wu, H., Pots, B., Hollenberg, J., Meerhoff, R., 1987. Flow pattern transitions in two-
Simmons, M.J., Hanratty, T.J., 2001. Transition from stratified to intermittent flows phase gas/condensate flow at high pressure in an 8-inch horizontal pipe. In:
in small angle upflows. Int. J. Multiph. Flow 27, 599–616. Proc. BHRA Conf., The Hague, The Netherlands, pp. 13–21.
Smith, I.E., 2017. A 7-field Lagrangian slug capturing and slug tracking model with Xiao, J., Shonham, O., Brill, J., 1990. A comprehensive mechanistic model for two-
higher order methods. phase flow in pipelines. In: SPE Annual Technical Conference and Exhibition.
Soleimani, A., Al-Sarkhi, A., Hanratty, T.J., 2002. Effect of drag-reducing polymers on Society of Petroleum Engineers.
pseudo-slugs––interfacial drag and transition to slug flow. Int. J. Multiph. Flow Yan, K., Che, D., 2010. A coupled model for simulation of the gas–liquid two-phase
28, 1911–1927. flow with complex flow patterns. Int. J. Multiph. Flow 36, 333–348.
Soleimani, A., Hanratty, T., 2003. Critical liquid flows for the transition from the Yu, Z., Hemminger, O., Fan, L.-S., 2007. Experiment and lattice Boltzmann
pseudo-slug and stratified patterns to slug flow. Int. J. Multiph. Flow 29, 51–67. simulation of two-phase gas–liquid flows in microchannels. Chem. Eng. Sci.
Štrubelj, L., Tiselj, I., 2011. Two-fluid model with interface sharpening. Int. J. Numer. 62, 7172–7183.
Meth. Eng. 85, 575–590. Zabaras, G., 1999. Prediction of slug frequency for gas-liquid flows. In: SPE Annual
Sussman, M., Smereka, P., Osher, S., 1994. A level set approach for computing Technical Conference and Exhibition. Society of Petroleum Engineers.
solutions to incompressible two-phase flow. Zhang, H.-Q., Jayawardena, S.S., Redus, C.L., Brill, J.P., 2000. Slug dynamics in gas-
Taha, T., Cui, Z., 2004. Hydrodynamics of slug flow inside capillaries. Chem. Eng. Sci. liquid pipe flow. J. Energy Resour. Technol. 122, 14–21.
59, 1181–1190. Zhang, H.-Q., Wang, Q., Sarica, C., Brill, J.P., 2003. Unified model for gas-liquid pipe
Taitel, Y., 1987. Effect of gas expansion on slug length in long pipelines. Int. J. flow via slug dynamics-part 1: model development. J. Energy Resour. Technol.
Multiph. Flow 13, 629–637. 125, 266–273.
Taitel, Y., Barnea, D., 1990. Two-phase slug flow. Adv. Heat Transfer 20, 83–132. Zhao, Y., Lao, L., Yeung, H., 2015. Investigation and prediction of slug flow
Taitel, Y., Dukler, A., 1976. A model for predicting flow regime transitions in characteristics in highly viscous liquid and gas flows in horizontal pipes.
horizontal and near horizontal gas-liquid flow. AIChE J. 22, 47–55. Chem. Eng. Res. Des. 102, 124–137.
Taitel, Y., Dukler, A., 1977. A model for slug frequency during gas-liquid flow in Zhao, Y., Yeung, H., Lao, L., 2013. Slug frequency in high viscosity liquid and gas flow
horizontal and near horizontal pipes. Int. J. Multiph. Flow 3, 585–596. in horizontal pipes. In: 16th International Conference on Multiphase Production
Toms, B.A., 1948. Some observations on the flow of linear polymer solutions Technology. BHR Group.
through straight tubes at large Reynolds numbers. Proc. In. Cong. Rheol. 1948, Zheng, G., Brill, J., Taitel, Y., 1994. Slug flow behavior in a hilly terrain pipeline. Int. J.
135. Multiph. Flow 20, 63–79.
Ujang, P.M., 2003. Studies of Slug Initiation and Development in Two-Phase Gas-
Liquid Pipeline Flow. Imperial College London (University of London).
19

Chemical Engineering Science: Abdalellah O. Mohmmed, Hussain H. Al-Kayiem, A.B. Osman

Uploaded by

Copyright:

Available Formats

Chemical Engineering Science: Abdalellah O. Mohmmed, Hussain H. Al-Kayiem, A.B. Osman

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chemical Engineering Science: Abdalellah O. Mohmmed, Hussain H. Al-Kayiem, A.B. Osman

Uploaded by

Copyright:

Available Formats

Chemical Engineering Science 238 (2021) 116611

Contents lists available at ScienceDirect

Chemical Engineering Science

Investigations on the slug two-phase flow in horizontal pipes: Past,

Among the potential two-phase flow regimes in a horizontal

Slug Pocket Slug Liquid

Slug Liquid Slug Pocket Slug Liquid Slug

Author Diameter jL (m/s) jG (m/s) Mean Comments and findings

Authors Pipe Fluids Slug frequency Correlation

FrS ¼ pjmffiffiffiffiffi (10)$1

(Fetter, 1988) 82 air–water- mS ¼ 0:0175Fr1:37

jm ¼ 24:73 þ 0:00766e 9:9121HLe

(Manolis et al., 1995) – air–water mS ¼ 0:037Fr1:8

(Zabaras, 1999) 25–203 air–water h i1:2 h i

(Al-Safran, 2016) 25–203 air-oil mS ¼ EXPð1:51  17:04ðDH Þ þ 0:77LnðjL Þ  0:181LnðsÞ(31)

Authors ID Co C1 jL (m/s) ReL (103) Remark

jG Thus, the translational slug velocities can be estimated using Eq.

Fig. 4. Distribution parameter as a function of mixture velocity – WASP (Hale, 2001).

Ghiaasiaan, 2006, Qian and Lawal, 2006, Kashid et al., 2008,

flows. However, few attempts have been conducted to study the

You might also like

jm ¼ 24:73 þ 0:00766e 9:9121HLe

(Al-Safran, 2016) 25–203 air-oil mS ¼ EXPð1:51 17:04ðDH Þ þ 0:77LnðjL Þ 0:181LnðsÞ(31)