International Communications in Heat and Mass Transfer 118 (2020) 104841

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ichmt

Slug frequency for a gas-liquid plug flow: Review and development of a new T
correlation
Abderraouf Arabia,b, , Karim Raguic, Yacine Salhib, Abdelkader Filalid,e
⁎

a
SONATRACH, Direction Centrale Recherche et Développement, Avenue 1er Novembre, 35000 Boumerdes, Algeria
b
University of Sciences and Technology Houari Boumediene USTHB, Physics' Faculty Laboratory of Theoretical and Applied Fluid Mechanics, LMFTA, BP 32 El Alia 16111
Bab Ezzouar, Algiers, Algeria
c
Process and Mechanical Engineering Department, USTHB University of Sciences and Technology, BP 32 El Alia 16111, Bab Ezzouar, Algiers, Algeria
d
Ecole Nationale Polytechnique de Constantine, BP 75, A, 25000 Nouvelle ville RP, Algeria
e
Chemical Engineering Department, Imperial College London, London, South Kensington London SW7 2AZ, UK

ARTICLE INFO ABSTRACT

Keywords: Slug frequency is one of the most important parameters for designing industrial installations where intermittent
Two-phase flow gas-liquid flows could be found. Thus, the literature is found full of various experimental investigations and
Plug flow developed empirical correlations to predict such parameter. Nevertheless, the effect of intermittent flow sub-
Plug frequency regimes on slug frequencies remains an open issue, especially in the case of plug flow. Hence, the present work is
Experimental data analysis
devoted to providing an update review and analysis of the slug frequency in plug flow case. For this purpose,
Empirical data
data from previous experiments have been collected and deeply examined. It was found that the increase in
Complementary correlation
liquid superficial velocity and the decrease in pipe diameter could induce an increase of plug frequency. The
comparison between experimental data and the predictions of twenty available empirical models showed that
the correlations based on Strouhal number and input liquid fraction are the best ones that predict well the
experimental plug frequency. However, each correlation is found to be valid for a limited range close to the
data's conditions that were used for its development. As such, and by considering an extended range of the whole
available experimental database, a new empirical correlation has been proposed.

1. Introduction when several phases are considered. Such type of flows includes gas-
liquid-liquid [37], gas-liquid-solid [38], liquid/liquid [39], liquid/solid
In a context characterized by energy savings and cost reduction, due [40], gas/solid [41] or gas-liquid flow [42,43]. The latter has been
to the growing industrial development in the world [1], theoretical widely encountered in various industries such as petroleum pipelines,
[2,3], numerical simulations [4–17] and experimental investigations nuclear power plants, power plants, refrigeration industries, heat ex-
[18–28] have been carried out in order to advance in the understanding changers, petrochemical and chemical process [44].
of several phenomena allied to fluid mechanics. These studies have The main particularity of this kind of flows is the notion of flow
allowed an improvement, optimization and innovate of some processes patterns which refers to the arrangement of phases in space and time
in which fluids are used. Indeed, some experimental approach carried inside the pipes. These flow patterns were regrouped in three main
out at laboratory scale, especially for the study of the influencing families: dispersed flow, separated flow and intermittent flow [45]. The
parameters individually [23,29] and for the development of new em- latter remains a relatively challenging topic and incites more com-
pirical correlations [30–32], which act as predictive models to deal plexity with a lack of complete understanding in some applications
with the lack of mathematical models [33]. The results obtained by [46]. Thereby, a part of this type of flows will be the subject of the
carrying out ‘controlled’ experiments were used as a benchmark data present paper.
for the validation of those obtained with numerical simulations The intermittent flow is to be present as an alternating flow of two
[34–36]. structures: liquid slugs (also known as liquid piston) and large gas
Beyond that aspect, multiphase flows are aroused to go further in pockets. For an industrial, intermittent flows could cause several pro-
the fluid mechanics area in the aim to deal with flow characterizations blems, including flow induced corrosion (FIC) [47], flow accelerated

⁎
Corresponding author at: SONATRACH, Direction Centrale Recherche et Développement, Avenue 1er Novembre, 35000 Boumerdes, Algeria.
E-mail addresses: [email protected] (A. Arabi), [email protected] (K. Ragui), [email protected] (Y. Salhi), [email protected] (A. Filali).

https://doi.org/10.1016/j.icheatmasstransfer.2020.104841

Available online 02 September 2020

0735-1933/ © 2020 Elsevier Ltd. All rights reserved.
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

Nomenclature Greek letters

AARD Average Absolute Relative Error [−] θ Pipe Inclination [°]

D Pipe Diameter [m] Θ Dimensionless function [−]
f Frequency [Hz] λ Volume fraction [−]
Fr =( V
gD ) Froude number [−] ρ
Φ
Density [Kg.m−3]
Dimensionless function [−]
g Gravity [m.s−2]
Ψ Dimensionless function [−]
ID Inner Diameter [m]
3 μ Viscosity [Pa.s]
D 2 L( L G )g
Nf = µL
Inverse Viscosity number [−]
Subscripts
S Slip factor [−]
Re = ( VD
µ ) Reynolds number [−] Cal Calculated
G Gas
St ( )
fV
=D Strouhal number [−]
L Liquid
SDRE Standard deviation of Relative Error [−]
M Mixture
V Velocity [m.s−1]
Mea Measured
X = L VSL
Lockhart-Martinelli parameter [−] S Superficial
G VSG

corrosion (FAC) [48], compressor overload and overflow liquid through the pressure drop depends directly on the slug frequency [74]. More-
downstream installations [49], to name a few. The alternation contacts over, some mechanical models use the slug frequency as a closure re-
between the gas-liquid phases and the pipe boundaries (e.g. inner wall) lationship [75]. Dong et al. [76] showed a strong relationship between
may result in significant variation in the pipe temperature, pressure the slug frequency and the heat transfer coefficient. For the afore-
gradient and vibrations [50], which could yield the pipe fatigue and its mentioned reasons, an accurate estimation of slug frequency is of pri-
destruction. Considering that this kind of flow is the most encountered mary interest to design and control various industrial processes where
flow pattern. Hence, dealing with its characteristics and controlling its an intermittent flow is encountered.
mechanisms remains a challenging task for engineers. Due to the limitations of theoretical models for the estimation of
Through horizontal and inclined pipes, the intermittent flow can be slug frequency, various empirical models have been widely adopted
presented in two different aspects: plug and slug flow. The distinction [60]. Most of these correlations were developed from measurements
between these two sub-regimes is still the subject of extensive debate in collected from experiments carried out at a laboratory scale, using
the scientific community. Indeed, for certain authors [51–54], the principally water and air as working fluids. It is worth denoting that
presence of air bubbles inside the liquid slugs could differentiate these each researcher, according to his own experimental installation, has
regimes (i.e. slug flow is characterized by aerated slugs, while in plug flow, developed his correlation according to certain parameters (like super-
there are no air bubbles inside the liquid piston). Other researchers [55–58] ficial fluids velocities, fluids properties and pipe diameter), while neglecting
used the interface-nose shape of the liquid slug/tail of the gas pocket as others. As a result, and based on the different used assumptions, a wide
the main criterion to distinguish between such two regimes. variety of correlations currently exists while each of them is found valid
When most works agree with the fact that the plug/slug flow tran- only under specific conditions close to those used for its elaboration
sition can occur for a certain gas superficial velocity [51,55,59,60], [77].
some studies have shown that the plug/slug transition depends on the If the influence of geometric parameters such as the pipe diameter
flow rates of both phases. Barnea & Brauner [61] have established a [78–81], fluids properties [80,81], flow parameters [60,67,82,83] and
criterion for the plug/slug flow transition, which gives: VSL = 1.5 - VSG, the validity of existing slug frequency correlations [77,84–86], have
where VSL and VSG are the superficial liquid and gas velocities, re- been widely studied in the case of slug flow, the same is not true when
spectively. This transition criterion was derived from the bubbly/in- talking about the plug regime. Performing a review of various studies
termittent flow theoretical transition criterion of Taitel & Dukler [62]. for different intermittent flow parameters, including the slug frequency,
Bertola [63] used the signal fluctuations obtained using a single- Thaker & Banerjee [52] found that numerous researchers have ne-
fiber optical probe in order to examine the plug/slug transition in a glected the plug flow in the development of their correlations. Owing to
horizontal pipe with an internal diameter of 80 mm. By representing the presence of large interfacial area for mass, momentum and energy
the obtained parameters from signals (partial derivative of Probability transfer between gas bubbles and liquid slugs [87], as well as the hy-
Density function (PDF) and the Power Spectral Density (PSD)) as a drodynamic interactions between gas bubbles (e.g. implosion, coales-
function of the gas velocities for various superficial liquid velocity, cence and agglomeration [88]) inside liquid slugs, the hydrodynamic of
Bertola [63] denoted a discontinuity in the evolution of PDF and PSD slug flow is much different from that of plug flow. Subsequently, the
parameters, which indicates a transition from plug to slug flow. The developed correlations for the slug regime cannot be directly extend-
transition between the two regimes is found as far as VSL = 5.1–4.7 VSG. able to the plug flow [52]. By comparing its data, collected on a 30 mm
Recently, Wang et al. [64] used an ultrasound Doppler velocimetry to pipe-diameter, with predictions of several existing models, Arabi et al.
measure the liquid velocity for plug and slug flow. Using the ratio of [60] found that the most commonly developed models for the predic-
liquid velocity between the bubble area and the slug body, the authors tion of slug frequency have better correlates to the obtained data for the
proposed another criterion to distinguish plug to slug flows, as slug flow regime in comparison with the plug flow regime. Therefrom,
VSL = 2.14 VSG – 0.22. Arabi et al. [60] suggested the development of a specific correlation for
The intermittent flows are characterized by specific parameters, each sub-regime.
such as the slug frequency [65,66], the slug length [54,67,68], the slug From this literature review, it appears that less attention was de-
translational velocity [69–71] and the liquid holdup in slug liquid voted to the slug frequency in the case of a plug flow (also known as plug
[72,73]. Thus slug frequency is defined as the number of liquid slugs frequency), compared to the slug flow case. Due to the difference be-
passing through a point in the pipe during one second. The slug fre- tween plug and slug flows, it is necessary to perform a comprehensive
quency is a key parameter when studying the intermittent flow. In fact, study that focuses exclusively on the slug frequency's behavior for this

2
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

regime. Consequently, the present work draws a deep understanding of

the plug frequency, by initially reviewing existing correlations devel-
oped for the predictions of slug frequency. Afterward, data of plug
frequency are collected from various published works in order to
highlight the most relevant parameters that could influence the plug
frequency. The collected data are compared to various available cor-
relations in order to survey their applicability for such specific regime.
To do so, the absolute average relative error as well as the standard
deviation of the relative error are used. As conclusion, a new correla-
tion covering a wide range of database available in the literature for a
plug flow is suggested. The correlation is based on gas based Strouhal
number and input liquid fraction and provides a good prediction of the
plug flow.

2. Review of slug frequency correlations

As a beginning, and by analyzing the experimental data of Hubbard

[89], Gregory & Scott [82] found that the slug frequency (f) is pro-
portional to the superficial velocity of the liquid phase, and inversely
proportional to the pipe diameter. Then, Gregory & Scott [82] proposed
the first correlation to predict slug frequency, as follows:
1.2
VSL 19.75
f = 0.0226 + VM
g D VM (1)

where VM, is the mixture velocity defined as:

VM = VSG + VSL (2)

Greskovitch & Shrier [74] developed an empirical correlation (Eq.3)

based on the one provided by Gregory & Scott [82]. These Authors
recommend the use of their correlation for large pipe diameters, as the
latter takes into account the diameter effect much better than that de-
veloped by Gregory and Scott [82].
1.2
VSL 2.02 V2
f = 0.0226 + m
VM D gD (3)

A few years later, Heywood & Richardson [90] adopted the Gamma Fig. 1. Representation of data points in the maps of Mandhane et al. [110] and
Ray Technique to measure the slug frequency on a 42 mm ID pipe with comparison with transition lines of Barnea & Brauner [61], Bertola [63] and
an air-water mixture. As a result, the Gregory and Scott's correlation Wang et al. [64]. (a) Air-water data. (b) air-oil and air‑silicone oil data.
was modified as follows:
1.02 The proposed correlation (Eq.6) gives better results with data collected
VSL 2.02 V2 from pressurized systems.
f = 0.0434 + m
VM D gD (4) 1.8
2
VSL 25 + VM
f = 0.0037
Nydal et al. [91] found that for high values of the liquid superficial gD VM (6)
velocity, the effect of the gas superficial velocity is small. Consequently,
the proposed correlation (Eq. 5) was based only on the liquid superficial Sanchez Silva et al. [93] proposed a correlation from their experi-
velocity and the pipe-diameter. mental results obtained with a 44.4 mm ID pipe. This correlation is
found to be closely consistent with the correlation of Greskovtich &
(VSL + 1.5) 2 Shrier [74].
f = 0.088
gD (5)
1.2879
VSL 20.308
Manolis et al. [92] measured the slug frequency for various oper- fs = 0.0148 + 1.3436 Vm
gD Vm (7)
ating pressures. They modified the correlation of Gregory & Scott [82].

Table 1
Details of the collected experimental data under various conditions.
Authors Kokal [106] Bertola & Cafaro [107] Thaker & Banerjee Conte et al. [57] Abulkadir et al. [108] Arabi et al. [60]
[52,88]

ID [mm] 25.8, 51.2 and 76.3 80 25 26 67 30

Working fluids Air/oil Air/water Air/water Air/water Air/ silicone oil Air/water
Gas superficial velocity range (VSG) [m/s] 0.046–0.718 0.151–0.592 0.484–0.726 0.2–0.5 0.05–0.058 0.436–0.786
Liquid superficial velocity range (VSL) [m/s] 0.03–0.61 0.6–2.0 0.194–0.975 0.3–0.7 0.095–0.38 0.141–0.778
Slug frequency range (f) [Hz] 0.025–0.747 0.147–2.43 0.332–3.187 0.636–3.62 0.136–0.595 0.336–2.106
Input liquid fraction range (λL) 0.061–0.930 0.503–0.903 0.211–0.668 0.429–0.714 0.621–0.884 0.153–0.641
Gas based Strouhal number range (StG) 0.004–0.521 0.040–0.879 0.011–0.165 0.041–0.314 0.182–0.797 0.013–0.145
Data points 42 16 16 5 6 20

3
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

Fig. 2. Influence of VSL and VSG on plug frequency. Data of Thaker & Banerjee Fig. 5. Gas based Strouhal number vs Lockhart-Martinelli parameter.
[52,88].

Fig. 6. Gas based Strouhal number vs mixture Froude number.

Fig. 3. Influence of VSG on plug frequency. Data of Bertola & Cafaro [107].
1.2
VSL 19.75
fs = 0.0226 + VM (0.836 + 2.75sin0.25 )
gD VM (8)

After measuring the slug frequencies, using the void fraction signal
on 40 and 60 mm ID pipes, Fossa et al. [94] proved that their experi-
mental data obtained on these pipe diameters, side by side with the
existing data from the literature, could be correlated using the gas
based Strouhal number (StG) (Eq.10) [95] and the inlet liquid fraction
(λL) (Eq.11). The obtained adjustments clear-up the following correla-
tion:
0.044 L
StG =
1 1.71 L + 0.70 2
L (9)

fD
StG =
Vsg (10)

VSL
L =
VSL + VSG (11)
Fig. 4. Influence of Pipe diameter on plug frequency. Data of Thaker & Banerjee
[52,88] and Bertola & Cafaro [107]. In small diameter pipes, the behavior of slug flow is found different
compared to the large ones. Such a difference is mainly due to the in-
Comparing different models of slug frequency with experimental fluence of the surface tension. Based on this evidence, Wren et al. [96]
results obtained on horizontal and inclined pipes, Zabaras [80] found modified the Gregory & Scott's correlation using collected data on a
that the existing models did not give good results for the inclined pipes' 5 mm diameter pipe. The proposed correlation is given as:
cases. Therefore, he proposed an empirical correlation that takes into 0.32
VSL 9.63
account the inclination effect. This correlation (Eq. 8) is based on the f = 1.278 + 1.2 VM
correlation of Gregory & Scott [82].
gD VM (12)

Using experimental data, obtained on pipes of an inner diameter of

4
 A. Arabi, et al.

Table 2
Performance evaluation of the existing correlations against the experimental data.⁎
Correlation Thaker & Banerjee Kokal [106] (25.6 mm) Conte et al. [57] Arabi et al. [60] Kokal [106] (51.8 mm) Abdulkadir et al. Kokal [106] (76.3 mm) Bertola & Cafaro All database
[52,88] (25 mm) (26 mm) (30 mm) [108] (67 mm) [107] (80 mm)

16 data 14 data 5 data 20 data 21 data 6 data 7 data 16 data 105 data

AARE SDRE AARE SDRE AARE SDRE AARE SDRE AARE SDRE AARE SDRE AARE SDRE AARE SDRE AARE SDRE

Gregory & Scott 75,05 60,46 1380,77 1003,50 54,20 58,30 22,80 33,48 485,45 345,96 201,21 139,19 95,44 86,52 59,21 108,93 326,43 605,29
[82]
Greskovich & Shrier 112,05 83,55 1746,52 1246,42 69,61 71,72 40,18 43,03 565,76 370,45 221,49 157,13 136,18 130,22 64,48 118,85 405,62 747,01
[74]
Heywood & 85,54 72,80 1516,75 1091,83 54,68 62,74 29,87 38,83 558,85 367,34 234,29 163,69 150,12 137,68 69,77 127,19 369,36 660,44
Richardson
[90]
Nydal et al. [91] 71,79 79,26 1208,13 939,40 40,33 45,40 35,94 47,28 297,55 322,36 43,68 61,43 155,43 234,26 45,24 69,76 260,05 547,89
Manolis et al. [92] 249,77 82,53 2710,84 2039,03 245,96 110,68 110,09 69,62 869,35 750,95 386,73 187,06 107,83 84,29 84,91 148,84 648,28 1188,55
Silva et al. [93] 36,34 40,79 903,87 679,44 32,74 39,63 27,94 22,59 297,57 234,18 104,43 94,32 51,88 59,27 54,85 74,28 210,23 409,67

5
Zabaras [80] 52,54 50,56 1138,17 839,02 42,02 48,74 20,81 27,99 389,37 289,21 151,99 116,45 67,86 72,15 55,90 91,11 265,33 506,09
Fossa et al. [94]. 16,06 13,98 145,35 97,75 22,10 17,87 23,21 17,35 158,91 157,79 29,75 37,66 53,36 61,84 73,55 83,55 75,55 112,04
Wren et al. [96] 289,43 226,44 3968,40 2925,37 142,40 137,47 201,45 149,09 1867,83 1458,21 790,91 518,51 1195,67 1249,14 329,13 413,05 1167,00 1824,64
Wang et al. [83] 22,98 20,75 142,29 93,38 29,78 23,52 20,90 16,01 77,39 58,48 44,28 12,34 49,42 63,66 46,10 77,95 56,20 78,03
Gokcal et al. [97] 86,02 3,30 31,74 36,12 91,75 2,73 91,96 1,58 58,15 14,39 91,33 1,94 81,26 12,40 95,40 2,66 76,03 35,77
Azzi et al. [98] 108,32 49,34 292,42 136,36 33,43 42,32 42,35 27,88 195,23 101,78 17,48 21,35 99,12 129,84 97,29 113,13 126,63 126,22
Schulkes [100] 131,50 164,66 1246,56 1421,99 62,63 46,97 84,32 129,45 232,18 465,47 89,91 7,62 276,48 427,94 69,41 39,01 285,87 707,01
Zhao et al. [78] 78,69 99,00 1313,64 1478,20 74,03 24,58 69,01 80,29 253,91 508,99 89,45 9,36 313,51 490,29 90,89 11,03 294,46 748,90
Vincesio et al. 25,76 11,05 170,16 140,00 16,03 18,01 23,78 25,50 218,50 230,10 57,63 54,90 51,12 67,60 128,05 108,60 101,82 151,71
[101]
Thaker & Banerjee 23,63 11,19 156,59 115,44 17,12 18,88 21,73 21,17 142,66 123,02 18,93 26,15 44,71 52,02 74,83 93,33 73,43 103,30
[52]
Al-Safran [102] 32,00 30,64 263,36 154,67 29,11 29,55 14,95 18,88 219,90 124,20 23,81 24,82 80,37 99,45 59,82 99,99 104,04 139,51
Abdulkadir et al. 59,64 54,52 197,19 128,47 43,84 40,45 20,88 27,01 79,31 92,63 67,84 13,54 97,36 132,47 60,34 66,07 76,87 107,06
[103]
Archibong-Eso 98,85 0,56 72,74 18,62 99,15 0,41 99,18 0,31 87,11 7,22 94,23 2,94 89,10 14,88 98,57 1,39 92,14 12,38
et al. [79]
Arabi et al. [60] 52,55 33,81 189,00 120,08 32,35 36,79 13,37 19,77 87,00 71,22 56,50 18,39 64,96 91,37 60,79 86,32 71,52 96,10

⁎
Bold numbers mean the best predictions for each database.
International Communications in Heat and Mass Transfer 118 (2020) 104841
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

Fig. 7. Comparison of the results predicted by the correlation of Fossa et al. Fig. 10. Comparison of the results predicted by the correlation of Arabi et al.
[94] with the experimental database. [60] with the experimental database.

Fig. 11. Gas based Strouhal number vs input liquid fraction.

Fig. 8. Comparison of the results predicted by the correlation of Wang et al.
[83] with the experimental database.
50.8 mm with oils (dynamic viscosity of 0.181 and 0.589 Pa.s) as a liquid
phase with different correlations available in the literature, Gokcal
et al. [97] found that existing correlations could not accurately predict
their obtained experimental results. Indeed, these correlations have
been developed using only water as a working liquid phase. Raison to
this, Gokcal et al. [97] developed a slug frequency's correlation which
takes into account the liquid-viscosity effect, as:
VSL
f = 2.623
N f0.612 D (14)
with Nf, is the inverse viscosity number, defined as:
3
D 2
L( L G )g
Nf =
µL (15)
The investigation of gas-liquid flow behavior through a micro-T-
junction was carried out by Azzi et al. [98] who represented a variety of
experimental data obtained from the literature (with pipe diameter
ranging from 5 to 150 mm) using the gas based Strouhal number (StG)
Fig. 9. Comparison of the results predicted by the correlation of Thaker &
according to the Lockhart-Martinelli parameter (X) (Eq. 16) (which was
Banerjee [52] with the experimental database.
introduced by Lockhart & Martinelli [99]). The obtained data is cor-
related using a linear adjustment provided by Eq. 17.
50 mm, and applying the same approach of Fossa et al. [94], Wang et al.
[83] proposed the following correlation: L VSL
X=
G VSG (16)
0.05 L
StG =
1 1.675 L + 0.768 2
L (13) StG = 0.0039 X (17)
Comparing the experimental data of an inner’ pipe-diameter of Based on a physics intuition, Schulkes [100] argued that the slug

6
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

76,28
SDRE
All database

105 data

AARE

50,44
Bertola & Cafaro [107]

76,80
SDRE
(80 mm)

16 data

AARE

39,86
16,49
SDRE
Kokal [106]
(76.3 mm)

7 data

AARE

26,84

Fig. 12. Comparison of the results predicted by the proposed correlation with
the experimental database.
Abdulkadir et al. [108]

Table 4
46,10
SDRE

Performance evaluation of the existing correlations and the developed corre-

lation against the experimental data of Zitouni et al. [54].
(67 mm)

Correlation Zitouni et al. [54] (40 mm)

6 data

AARE

41,34

5 data

AARE SDRE
76,71
SDRE

Gregory & Scott [82] 91,66 85,02

Kokal [106]
(51.8 mm)

Greskovich & Shrier [74] 138,09 109,88

21 data

Heywood & Richardson [90] 123,64 103,42

AARE

80,32

Nydal et al. [91] 99,72 100,47

Manolis et al. [92] 169,49 105,05
Silva et al. [93] 44,03 58,53
Zabaras [80] 61,25 71,12
88,41
SDRE

Fossa et al. [94]. 39,09 51,77

Arabi et al. [60]

Wren et al. [96] 584,74 358,04

Wang et al. [83] 46,58 57,64
(30 mm)

Gokcal et al. [97] 88,74 4,77

20 data

115,89
AARE

Azzi et al. [98] 158,08 108,32

Schulkes [100] 173,38 157,66
Zhao et al. [78] 57,78 71,29
Vincesio et al. [101] 38,22 49,06
Performance evaluation of the developed correlation against the experimental data.

14,35
SDRE

Thaker & Banerjee [52] 39,90 51,22

Conte et al. [57]

Al-Safran [102] 92,54 80,99

Abdulkadir et al. [103] 103,16 89,96
(26 mm)

Archibong-Eso et al. [79] 98,34 0,80

5 data

AARE

32,59

Arabi et al. [60] 88,37 75,77

Proposed correlation (Eq. 40) 35,70 45,59
15,89
SDRE

frequency depends on eight (08) parameters, which are summarized as:

the superficial fluid velocities (VSG, VSL), acceleration gravity (g), pipe's
Kokal [106]
(25.6 mm)

diameter (D), dynamic liquid viscosity (μL), fluid densities (ρG, ρL) and
14 data

AARE

29,07

the pipe's inclination (θ). As such, Schulkes [100] presented his general
empirical correlation, as follows:
VM
f= ( G) (ReSL) ( , Fr )
(18)
11,69
SDRE

D
Thaker & Banerjee
[52,88] (25 mm)

where:
( G) = 0.016 G (2 +3 G) (19)
16 data

AARE

9,31

12.1 ReSL0.37 for ReSL < 4000

(ReSL) =
Proposed correlation

1 for ReSL 4000 (20)

2
1+ sin( ) | | for | | 0.17
FrSL
( , FrSL) =
Table 3

1.8 2)
(0.6 + 2 for | | > 0.17
FrSL (21)

7
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

=
VSG new correlation, as follows:
G
VSL + VSG (22)
L
VSL D lnf = ln
ReSL = L N fn (34)
µL (23)
where:
VSL
FrSL = (0.75, 0.55) for ReSG 2500
gDcos (24) ( , n) =
(0.35, 1.02) for ReSG > 2500 (35)
Zhao et al. [78] investigated the slug frequency for air-oil mixture
on two pipes diameters (26 mm and 44 mm). The dynamic viscosity At the latest, Arabi et al. [60] generated a database of slug fre-
values of the used oils have been taken as 1.0, 3.5, 6.0 and 7.0 Pa.s. quency for various intermittent flow sub-regimes (Plug flow, Slug flow
After evaluating the viscosity effects, Zhao et al. [78] proposed the and Highly Aerated Slug flow) using a 30 mm ID pipe. After a deep
following correlation: comparison between their experimental data and those of a large
variety of available correlations, Arabi et al. [60] displayed that the
VM predictive levels of the correlations depend on the nature of the sub-
10.836 ( ) ReSL0.337 for ReSG 4000
f= D for ReSL < 4000 regime. The authors concluded that there is no ultimate correlation
VM
6.40 ( ) ReSL0.141 for ReSG > 4000 available and reliable for the three sub-regimes. For this reason, Arabi
D (25) et al. [60] provided a complement correlation that takes into account
where: the nature of the sub-regimes. Each sub-regime data has been correlated
separately, using the gas Strouhal number and input liquid fraction.
VSG D
ReSG = G
Regarding the case of plug regime, this correlation is given as follows:
µG (26)
0.06327 L
StG =
( ) = 0.016 G (2 3 G) (27) 1 1.855 L + 1.163 2
L (36)
Using a 25.8 mm ID pipe, Vicensio et al. [101] accomplished the Arabi et al. [60] used Thaker & Banerjee data [88] as a validation
open literature by their empirical correlation, given as follows: databank. Hence, a great concordance between Arabi et al. [60] cor-
0.45 relation and Thaker & Banerjee [88] data was highlighted. It is worth
VSL VSL
f = 0.08 denoting that Thaker & Banerjee data have been used only for a vali-
D VSG (28) dation purpose and have not been used for developing the correlation.
Thaker & Banerjee [52] measured the slug frequency within a
25 mm ID pipe. Then they compared the obtained results for the plug 3. Available experimental data for slug frequency in plug flow
and slug regimes with the values given by the correlations of Fossa et al.
[94] and Wang et al. [83], Thaker & Banerjee [52] observed that these For the present investigation, a database collected from seven (7)
two models could be well adapted with their slug flow results, but these references and obtained from six (6) experimental devices are deeply
latter underestimate the obtained plug flow data. Consequently, the analyzed and compared. For the author's best knowledge, these slug
authors used their data obtained for both regimes to propose a new frequency data are the only experimental investigations that have been
correlation, (given by Eq. 29) for the prediction of slug frequency. carried out for the plug flow case. Details of the experimental condi-
Unlike the correlations of Fossa et al. [94] and Wang et al. [83], such an tions of these studies are summarized in Table 1. It is to note that all
empirical correlation is based on the mixture velocity (StM) (Eq. 30). these experiments were conducted using water and air as working
fluids, except those of Kokal [106] and Abdulkadir et al. [108] for
0.036 L
StM = which, oil and silicone oil have been used as a liquid phase, respec-
1 1.432 L + 0.739 2
(29)
L
tively. The inner pipe diameter is within the range of 25 and 80 mm, so
fD these data concern the small and medium diameter pipe, according to
StM = the summary classification given recently by Abdulkadir et al. [109].
VM (30)
As displayed in Fig. 1, these data were plotted on the flow pattern
After having explained the hazardous nature of the slug frequency map of Mandhane et al. [110]. The latter was considered as the re-
parameter, Al-Safran [102] proposed a probabilistic model for the ference flow pattern map when dealing with horizontal configuration.
prediction of this parameter. This correlation is based on Poisson's law. The theoretical transition criteria of Barnea & Brauner [61] as well as
lnfs = [1.51 17.04D + 0.77 Ln (VSL) 0.181 Ln (S )] (31) the experimental one of Bertola [63] and Wang et al. [64] are also
represented through Fig.1. Data obtained with air-water and air-oil and
VG µ
0.01
VSG
1.5 silicone oil as well, were represented separately in Fig.1(a) and
S= = 1 + 1.95 L Fig.1(b), respectively. Regarding Fig.1(a), the experimental points of all
VL µ0 VSL (32)
studies are found to be located in the plug flow area. A slight shift is
With μ0= 1 Pa.s, as a reference value of viscosity. observed for the data of Arabi et al. [60] with a transition line between
Abdulkadir et al. [103] correlated his data obtained on a 67 mm the stratified flow/plug flow of the map of Mandhane et al. [110]. It can
pipe with a silicone oil-air mixture using the gas based Strouhal number be noticed that Bertola's criteria for the transition plug to slug flow
and the input liquid fraction. Contrary to the correlations of Fossa et al. could predict the entire database. However, remaining criterions could
[94] and Wang et al. [83], a power function was used to fit the data as not provide satisfactory results.
follows: Concerning the data of Kokal [106] and Abdulkadir et al. [108], (i.e.
Fig. (b)), these latter have been situated in both the plug and the
StG = 0.184 1.274
(33)
L
stratified zones. Indeed, and compared to air-water mixture, a lesser
Later, Archibong-Eso et al. [79] examined the impact of the gas amount of liquid superficial velocity is required to generate intermittent
superficial velocity and the liquid viscosity on slug frequency using flow with air-oil system. This behavior was mainly denoted and ex-
their data collected on air-oil mixtures into pipes with inner diameters plained theoretically and experimentally by Andritsos et al. [111].
of about 25.4 and 76.4 mm. Based on their data and those of Gokcal Following, the transition criterion of Barnea & Brauner [61] and those
et al. [97], Farsetti et al. [104] and Baba et al. [105], they proposed a of Bertola [63] have well predicted such data.

8
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

4. Discussion absolute average relative error (AARE) (as shown in Eq.38), which is
used to quantify the difference between the experimental values (fmea)
4.1. Experimental data analysis and the correlation values (fcal). The latter is the standard deviation of
the relative error (SDRE) as given by Eq.39, which expresses the dis-
In order to clear-up the effect of superficial velocities of the fluid persion of relative errors between the two values (i.e. the calculated and
phases on the plug frequency, the variation of the latter for Thaker & the measured).
Banerjee data [52,88] is displayed in Fig.2. The increase in the flow rate N
1 fcal fmea
of the liquid phase leads to an increase in the frequency of the liquid AARE = × 100
slugs' passage. Similar behavior was observed in the slug flow case of N I=1
fmea (38)
many experiments [52,60,67,70,83]. Indeed, the increase in the liquid
superficial velocity leads to a rise in the liquid height within the pipe, N 2
1 fcal fmea
SDRE = × 100
which provokes a generation of additional liquid slugs [112]. Mean- N fmea
I=1 (39)
while, an increase in the gas superficial velocity results a decrease in the
plug frequency. However, this influence is found to be very weak with N as a data number.
compared to that of the liquid phase. The corresponding results of the parameters AARE and SDRE, given
The effect of the gas superficial velocity on the plug frequency was in %, for each correlation and each database are summarized in Table.2.
also investigated using the data of Bertola & Cafaro [107], in which the Based on the analysis of the two statistical parameters, it can be seen
results are displayed in Fig.3. Hence, it is observed that the flow rate of that:
the gas phase provides minor effects on the plug frequency. Regarding
the slug flow, various studies [82,90,113,114] found that the increase - A large difference between the statistical values of AARE for the
in gas superficial velocity leads to a decrease in the slug frequency up to different correlations is obtained. Such a difference is due to the
VSG ~ 4–6 m/s. Above such a critical value, an increase in the gas su- wide variety of the existing correlations for the estimation of the
perficial velocity could lead to an increase in the number of formed slug frequency.
slugs. This can be explained by the slug formation mechanisms. Indeed, - The correlation of Wang et al. [83], with AARE and SDRE of about
for low superficial gas velocity, the slugs were generated by the wave 56.20% and 78.03%, respectively, is the one that best predicted the
growth mechanism [65], when the rise in gas superficial velocity has an experimental data for the plug regime. The correlations of Arabi
inverse effect on the formation of slugs. Such a critical influence is et al. [60], Thaker & Banerjee [52] and Fossa et al. [94] follow with
found to be relatively weak and depends on the liquid superficial ve- AARE coefficients equals to 71.52%, 73.43% and 75.55%, respec-
locities as denoted by Lin et al. [114]. tively. These four correlations have the particularity to use Strouhal
In order to investigate the influence of the pipe diameter, the data of numbers and input liquid fraction.
Thaker and Banerjee [52,88] and Bertola & Cafaro [107] is represented - By taking each database separately, the Thaker & Banerjee data
as a function of the mixture velocity (VM) in Fig.4. These data have been [52,88] is better predicted with the correlation of Fossa et al. [94]
obtained with 25 and 80 mm ID pipe, respectively. Then, it is clearly (AARE is about 16.06%).
shown that the increase in the pipe diameter results in a decrease in the - The correlations of Vincensio et al. [101] and Thaker & Banerjee
plug frequency. Indeed, a large pipe diameter requires more liquid flow [52] predicted well the data collected on a 26 mm diameter pipe of
rates to increase the liquid height which will lead, in turn, to the for- Conte et al. [57] with AARE equals to 16.03% and 17.12%, re-
mation of additional liquid slugs. A similar behavior was observed for a spectively. This can be explained by the fact that these correlations
slug flow by several authors such as Gregory & Scott [82] and Archi- were developed from a data of pipe diameter of about 25.8 mm and
bong-Eso et al. [79]. 25 mm, respectively, which is so close to that used by Conte et al.
The gathered experimental database is represented through Fig. 5, [57].
using the gas based Strouhal number (StG) as a function of Lockhart- - Experimental data of Arabi et al. [60] were correlated very well
Martinelli parameter (X). Indeed, the obtained linear trend reveals that using their empirical model with AARE equal to 13.37%.
the use of the dimensionless gas based Strouhal number to express the - The data from Abulkadir et al. [108] were well predicted (in the
plug frequency is very interesting. Noted that the linear trend was al- following order) by the correlations of Azzi et al. [98], Thaker &
ready observed in the case of slug flow by several studies [60,98,103]. Banerjee [52] and Al-Safran [102], with values of AARE equal re-
Through Fig. 6, the dimensionless slug frequency (using the gas based spectively to 17.48%, 18.93% and 23.81%. However, this database,
Stouhal number) is represented as a function of the mixture Froude as well as those of Conte et al. [57] are composed of a small number
number (FrM) (i.e. Eq.37). Such a type of representation was proposed of data, which makes the interpretation of such a result much
by Arabi et al. [60] as a tool in order to distinct between the different complicated.
sub-regimes of intermittent flow. Hence, all data used in this study are - About the data of Kokal [106], the data obtained with 25.8 mm and
situated on the plug flow area, which confirms the validity of this map. 51.2 mm ID are well correlated by the correlation of Gokcal et al.
It is worth denoting that Arabi et al. [60] have developed this flow map [97] with AARE equal to 31.74% and 58.15%, respectively. With
based on their results and those of Thaker & Banerjee [88]. AARE = 49.4%, the data obtained with a larger diameter were well
predicted with the correlation of Wang et al. [83].
VM
FrM = - Finally, the correlation of Nydal et al. [91] and Wang et al. [83] are
gD (37) the correlations that best predicted the data of Bertola & Cafaro
[107] with AARE equal to 45.24% and 46.10%, respectively. Even
with a quite large SDRE value (69.76% and 77.95%, respectively).
4.2. Comparison with existent correlations
For a further investigation, the correlations that best predicted the
The selected experimental data were compared with twenty (20) experimental data, namely the models of Fossa et al. [94], Wang et al.
correlations for the prediction of slug frequency. As shown in section [83], Thaker & Banerjee [52] and Arabi et al. [60], are analyzed gra-
§2, these correlations are very different in the form and the parameters phically and the experimental data given by these models are shown in
used. Consequently, it is important to analyze the assessment of these Figs.7–8-9 and 10, respectively. The abscissa axis represents the mea-
correlations in the case of plug frequency. For the prediction of each sured frequency while the ordinate axis represents the calculated fre-
presented correlation, two parameters were adopted. The first one is the quency from each correlation. For this study, the maximum limit of the

9
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

deviation was set at about ± 30%. respectively, it is found that the proposed correlation outperforms the
Fig.7 demonstrates that Fossa et al. correlation [94] predicted well other slug frequency correlations evaluated.
the great majority of these empirical data. However, the data obtained These primary results should not hide the fact that this study was
for 51.2 and 80 mm pipe's diameters were overestimated. Regarding the carried out on a small database (of about 105 points), due to the low
correlation of Wang et al. [83] (Fig.8), the later have overestimated the number of studies available for the plug flow as a result of the neglected
experimental results in the case of small values of plug frequency, slug frequency in the case of plug flow. Besides, the whole data used
which correspond to the case of a weak superficial liquid velocity. In- were obtained on small to medium pipe diameters. Subsequently, future
deed, Wang et al. [83] used the data collected for a relatively high investigations will be based on an extended range of pipe diameters in
superficial liquid velocity (> 0.35 m/s) to build their correlation. For order to increase the database of slug frequency through the case of
the remaining part of the data, the correlation of Wang et al. [43] plug flow. These databases will be very helpful to provide a better
predicted well the plug frequency. The Thaker & Banerjee correlation consistent model for such a specific regime, and will serve as a basis to
[52] (Fig.9) estimated accurately the various data obtained using water validate the simulations obtained with Computational Fluid &
as a liquid phase, except the data of Bertola & Cafaro's [107]. This can Multiphase Flow Dynamics (CFD & CMFD).
be explained by the range of input liquid fraction (λL) of these data.
Indeed, Bertola & Cafaro [107] performed their experiments with great 5. Conclusions
values of superficial liquid velocities which results in great values of
input liquid fraction (0.503 < λL < 0.903), when Thaker and The intermittent gas-liquid flow could be divided into two distinct
Banerjee correlation [52] is valid for input liquid fractions less than regimes: plug and slug flow. Each regime can provide its particular
0.65. shape and hydrodynamics properties. Most of the experimental works
Due to the limited database obtained for one diameter when de- and empirical correlations developed for the prediction of the slug
veloped, the correlation of Arabi et al. [60] (i.e.Fig.10) predicted only frequency have been carried out by neglecting the plug flow. Thus,
the data that have been used for its development, in a satisfactory way. investigating the slug frequency in the case of a plug flow was the main
Following this analysis, it appears that none of the existing corre- objective of the present paper. A review of existing correlations for slug
lations can accurately predict the entire database of the open literature. frequency predictions was carried out. A databank from various in-
Indeed, each correlation is found to be valid in the range close to the vestigations was collected and deeply analyzed. Hence, the main con-
data that was used for its development, which is a bit critical somehow. clusions hold as follows:
The reason why the correlation of Arabi et al. [60] performed well the
plug frequency is due to the fact that this model was developed by - The comparison between the data with the different plug/slug
incorporating only the data from the plug flow, despite the small transition models shown that the empirical model of Bertola [63]
amount of data used in its development (20), Thus, and following this predicts well the entire data used in this study.
primarily results, it seems very crucial to provide a novel correlation - The plug frequency is a main function of the liquid superficial ve-
based exclusively on plug flow data. locity and the pipe diameter, when the gas superficial velocity has a
weak influence.
4.3. Development of a new correlation - The detailed analysis of predictions of twenty available correlations
proved that the correlations of Fossa et al. [94], Wang et al. [83],
In order to generate an outstanding correlation specific to the plug Thaker & Banerjee [52] and Arabi et al. [60] are the models which
flow and valid for the wide range of the above database, we reported best estimate the experimental data. These correlations have the
data with gas based Strouhal number as a function of the input liquid particularity of using the Strouhal number and input liquid fraction
fraction, as displayed in Fig. 11. The same fit function used by Fossa as relevant parameters. Meanwhile, none of these models can ac-
et al. [94], Wang et al. [83] and Arabi et al. [60] was adopted to cor- curately predict the whole database.
relate the data. Hence, the obtained correlation developed in the pre- - Based on these observations, a new correlation was proposed for the
sent study is defined as follows: estimation of plug frequency. The latter is based on gas based
Strouhal number and input liquid fraction. Compared to existing
0.03693 L
StG = correlations, this specific correlation has built using data available
1 1.865 L + 0.9118 2
(40)
L
in the open literature.
Following, the performance of the developed correlation against
each experimental data was evaluated using the statistical parameters, Although, it is important to underline that this complementary
as shown in Table 3. With the values of AARE and SDRE equal to correlation was provided on small data numbers, due to the low number
50.44% and 76.28%, respectively, the proposed correlation is found to of studies available for the plug flow. Thus, the generation of additional
be the correlation that best predicted the entire database, compared to data for this specific regime and for various experiment conditions,
the available ones (Table 2). Likewise, the proposed correlation is the especially with large pipe diameter and using different viscosities of
one that best predicted the data of Thaker & Banerjee [52,88], Bertola & liquid phase (where there is a lack of experiments), is strongly re-
Cafaro [107], and Kokal [106] data for pipe diameters of 25.6 mm and commended in future works. As such, the development of accurate
76.3 mm, respectively. models will be reached.
The calculated plug frequencies from the proposed correlation (i.e.
Eq.40) have been displayed as a function of the measured data, as Declaration of Competing Interest
shown in Fig. 12. Indeed, almost all data are found to be in the 30%
confidence interval, which is very encouraging, except for the Kokal's The authors declare that they have no known competing financial
data [106] which, as a reminder, used oil as a liquid phase. interests or personal relationships that could have appeared to influ-
In order to evaluate the validity of the proposed plug frequency ence the work reported in this paper.
correlation, we compared the predictions of the selected correlations as
well as the proposed correlation with the data obtained using air-water References
mixture on a 40 mm ID pipe by Zitouni et al. [54]. As a reminder, these
data were not used in the development of our correlation. The statistical [1] M.R. Samani, D. Toghraie, Removal of hexavalent chromium from water using
parameters calculated for each correlation were given in Table 4. With polyaniline/wood sawdust/poly ethylene glycol composite: an experimental
the values of AARE and SDRE equal to 35.70% and 45.59%, study, J. Environ. Health Sci. Eng. 17 (1) (2019) 53.

10
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

[2] A. Rabinovich, P. Bedrikovetsky, D.M. Tartakovsky, Analytical model for gravity [28] M. Rezaee, M. Khoshvaght-Aliabadi, A.A. Arani, S.H. Mazloumi, Enhanced heat
segregation of horizontal multiphase flow in porous media, Phys. Fluids 32 (4) transfer in pin fin heat sink working with nitrogen gas–water two-phase flow:
(2020) 046602. variable pin length and longitudinal pitch, J. Therm. Anal. Calorim. (2019) 1–27.
[3] D. Panici, G.A. de Almeida, A theoretical analysis of the fluid–solid interactions [29] M.H. Aghahadi, M. Niknejadi, D. Toghraie, An experimental study on the rheo-
governing the removal of woody debris jams from cylindrical bridge piers, J. Fluid logical behavior of hybrid tungsten oxide (WO3)-MWCNTs/engine oil Newtonian
Mech. 886 (2020). nanofluids, J. Mol. Struct. 1197 (2019) 497–507.
[4] D. Toghraie, A.R. Azimian, Molecular dynamics simulation of annular flow boiling [30] A. Afshari, M. Akbari, D. Toghraie, M.E. Yazdi, Experimental investigation of
in a microchannel with 70000 atoms, Int. J. Physical Math. Sci. 5 (6) (2011) rheological behavior of the hybrid nanofluid of MWCNT–alumina/water
831–837. (80%)–ethylene-glycol (20%), J. Therm. Anal. Calorim. 132 (2) (2018)
[5] D.T. Semiromi, A.R. Azimian, Molecular dynamics simulation of annular flow 1001–1015.
boiling with the modified Lennard-Jones potential function, Heat Mass Transf. 48 [31] B. Ruhani, P. Barnoon, D. Toghraie, Statistical investigation for developing a new
(1) (2012) 141–152. model for rheological behavior of silica–ethylene glycol/water hybrid Newtonian
[6] D. Toghraie, Numerical thermal analysis of water's boiling heat transfer based on a nanofluid using experimental data, Physica A: Stat. Mech. Appl. 525 (2019)
turbulent jet impingement on heated surface, Physica E: Low-Dimen. Syst. 616–627.
Nanostruct. 84 (2016) 454–465. [32] B. Ruhani, D. Toghraie, M. Hekmatifar, M. Hadian, Statistical investigation for
[7] L. Khezzar, A. Filali, O. Matar, Flow instability in a plane sudden expansion fol- developing a new model for rheological behavior of ZnO–ag (50%–50%)/water
lowed by a curved duct, Bull. Am. Phys. Soc. 63 (2018). hybrid Newtonian nanofluid using experimental data, Physica A: Stat. Mech. Appl.
[8] K. Ragui, A. Boutra, R. Bennacer, N. Labsi, Y.K. Benkahla, Correlating heat and 525 (2019) 741–751.
mass transfer coefficients for thermosolutal convection within a porous annulus of [33] S.U. Khalid, H. Babar, H.M. Ali, M.M. Janjua, M.A. Ali, Heat pipes: progress in
a circular shape: case of internal pollutants spreading, Heat Mass Transf. 54 (7) thermal performance enhancement for microelectronics, J. Therm. Anal. Calorim.
(2018) 2061–2078. (2020), https://doi.org/10.1007/s10973-020-09820-7.
[9] Y. Peng, M. Zarringhalam, M. Hajian, D. Toghraie, S.J. Tadi, M. Afrand, [34] K. Ragui, A. Boutra, R. Bennacer, Y.K. Benkahla, Progress on numerical simulation
Empowering the boiling condition of argon flow inside a rectangular microchannel of yield stress fluid flows (part I): correlating thermosolutal coefficients of
with suspending silver nanoparticles by using of molecular dynamics simulation, Bingham plastics within a porous annulus of a circular shape, Int. J. Heat Mass
J. Mol. Liq. 295 (2019) 111721. Transf. 126 (2018) 72–94.
[10] S. Rostami, H. Ahmadi-Danesh-Ashtiani, D. Toghraie, R. Fazaeli, A statistical [35] R. Azadbakhti, F. Pourfattah, A. Ahmadi, O.A. Akbari, D. Toghraie,
method for simulation of boiling flow inside a platinum microchannel, Physica A: Eulerian–Eulerian multi-phase RPI modeling of turbulent forced convective of
Stat. Mech. Appl. 123879 (2019). boiling flow inside the tube with porous medium, Int. J. Num. Methods Heat Fluid
[11] M. Zarringhalam, H. Ahmadi-Danesh-Ashtiani, D. Toghraie, R. Fazaeli, The effects Flow. 30 (5) (2019) 2739–2757.
of suspending copper nanoparticles into argon base fluid inside a microchannel [36] M.R. Sheykhi, M. Afrand, D. Toghraie, P. Talebizadehsardari, Numerical simula-
under boiling flow condition by using of molecular dynamic simulation, J. Mol. tion of critical heat flux in forced boiling of a flow in an inclined tube with dif-
Liq. 293 (2019) 111474. ferent angles, J. Therm. Anal. Calorim. 139 (4) (2020) 2859–2880.
[12] M. Zarringhalam, H. Ahmadi-Danesh-Ashtiani, D. Toghraie, R. Fazaeli, Molecular [37] Y. Liu, J. Yue, C. Xu, S. Zhao, C. Yao, G. Chen, Hydrodynamics and local mass
dynamic simulation to study the effects of roughness elements with cone geometry transfer characterization under gas–liquid–liquid slug flow in a rectangular mi-
on the boiling flow inside a microchannel, Int. J. Heat Mass Transf. 141 crochannel, AICHE J. 66 (2) (2020) e16805.
(2019) 1–8. [38] P. Sassi, Y. Stiriba, J. Lobera, V. Palero, J. Pallarès, Experimental Analysis of
[13] Y. Peng, M. Zarringhalam, A.A. Barzinjy, D. Toghraie, M. Afrand, Effects of surface Gas–Liquid–Solid Three-Phase Flows in Horizontal Pipelines, Flow, Turbulence
roughness with the spherical shape on the fluid flow of argon atoms flowing into and Combustion, 2020, pp. 1–20.
the microchannel, under boiling condition using molecular dynamic simulation, J. [39] A. Abdollahi, S.E. Norris, R.N. Sharma, Fluid flow and heat transfer of liquid-liquid
Mol. Liq. 297 (2020) 111650. Taylor flow in square microchannels, Appl. Thermal Eng. (2020) 115123.
[14] A.R. Ravangard, L. Momayez, M. Rashidi, Effects of geometry on simulation of [40] L. Piantanida, A.F. Payam, J. Zhong, K. Voïtchovsky, Nanoscale mapping of the
two-phase flow in microchannel with density and viscosity contrast, J. Therm. directional flow patterns at liquid-solid interfaces, Physical Rev. Appl. 13 (6)
Anal. Calorim. 139 (1) (2020) 427–440. (2020) 064003.
[15] A. Shahsavar, H.M. Ali, R.B. Mahani, P. Talebizadehsardari, Numerical study of [41] J. Wang, Continuum theory for dense gas-solid flow: a state-of-the-art review,
melting and solidification in a wavy double-pipe latent heat thermal energy sto- Chem. Eng. Sci. 215 (2020) 115428.
rage system, J. Therm. Anal. Calorim. (2020), https://doi.org/10.1007/s10973- [42] X.Y. Duan, F.B. Li, B. Ding, L. Gong, M.H. Xu, Heat transfer characteristics and
020-09864-9 (in press). field synergy analysis of gas–liquid two-phase flow in micro-channels, J. Therm.
[16] S. Kherroubi, K. Ragui, A. Bensaci, N. Labsi, A. Boutra, Y.K. Benkahla, Effect of the Anal. Calorim. (2019) 1–12.
second outlet location and the applied magnetic field within a ventilated cubic [43] M. Meribout, A. Abdelwahid, N. Ghendour, N. Kharoua, L. Khezzar, An ultrasonic-
cavity crossed by a nanofluid on mixed convection mode: best configurations, J. based multiphase flow composition meter, Measurement (2020) 108111.
Therm. Anal. Calorim. 139 (3) (2020) 2243–2264. [44] E.M. El-Said, M.M. Awad, M. Abdulaziz, A comprehensive review on pressurized
[17] S.R. Yan, N. Shirani, M. Zarringhalam, D. Toghraie, Q. Nguyen, A. Karimipour, thermal shock: predictive, preventive and safety issues, J. Therm. Anal. Calorim.
Prediction the boiling flow characteristics in rough and smooth microchannels (2020) 1–20.
using of molecular dynamics simulation: investigation the effects of boundary wall [45] A.E. Dukler, M.G. Hubbard, A model for gas-liquid slug flow in horizontal and near
temperatures, J. Mol. Liq. 112937 (2020). horizontal tubes, Ind. Eng. Chem. Fundam. 14 (4) (1975) 337–347.
[18] S. Mengjie, D. Chaobin, H. Eiji, Experimental investigation on the heat transfer [46] J.M. Godhavn, M.P. Fard, P.H. Fuchs, New slug control strategies, tuning rules and
characteristics of novel rectangle radial microchannel heat exchangers in two- experimental results, J. Process Control 15 (5) (2005) 547–557.
phase flow cooling system for data centers, J. Therm. Anal. Calorim. (2019) 1–13. [47] H. Zhang, H.Q. Lan, A review of internal corrosion mechanism and experimental
[19] W. He, B. Ruhani, D. Toghraie, N. Izadpanahi, N.N. Esfahani, A. Karimipour, study for pipelines based on multiphase flow, Corros. Rev. 35 (6) (2017) 425–444.
M. Afrand, Using of artificial neural networks (ANNs) to predict the thermal [48] S. Zhai, S. Pang, J. Shao, H. Zhu, X. Zhang, P. Liu, The FAC experimental research
conductivity of zinc oxide–silver (50%–50%)/water hybrid Newtonian nanofluid, on vapor-water two-phase flow piping system of CAP1400 secondary loop, MATEC
Int. Commun. Heat Mass Transf. 116 (2020) 104645. web of conferences, 207 EDP sciences, 2018, p. 04001.
[20] S. Rostami, D. Toghraie, M.A. Esfahani, M. Hekmatifar, N. Sina, Predict the [49] A. Arabi, Contribution à l’étude du comportement d’un écoulement diphasique
thermal conductivity of SiO 2/water–ethylene glycol (50: 50) hybrid nanofluid dans une conduite en présence d’une singularité. PhD thesis, USTHB, Algiers,
using artificial neural network, J. Therm. Anal. Calorim. (2020) 1–10. Algeria, 2019.
[21] S. Rostami, D. Toghraie, B. Shabani, N. Sina, P. Barnoon, Measurement of the [50] Rahmandhika A. Deendarlianto, A. Widyatama, O. Dinaryanto, A. Widyaparaga,
thermal conductivity of MWCNT-CuO/water hybrid nanofluid using artificial Indarto, Experimental study on the hydrodynamic behavior of gas-liquid air-water
neural networks (ANNs), J. Therm. Anal. Calorim. (2020) 1–9. two-phase flow near the transition to slug flow in horizontal pipes, Int. J. Heat
[22] Y. Zheng, H. Yang, M.A. Fazilati, D. Toghraie, H. Rahimi, M. Afrand, Experimental Mass Transf. 130 (2019) 187–203.
investigation of heat and moisture transfer performance of CaCl2/H2O-SiO2 na- [51] D. Barnea, O. Shoham, Y. Taitel, A.E. Dukler, Flow pattern transition for gas-liquid
nofluid in a gas–liquid microporous hollow fiber membrane contactor, Int. flow in horizontal and inclined pipes. Comparison of experimental data with
Commun. Heat Mass Transf. 113 (2020) 104533. theory, Int. J. Multiphase Flow 6 (3) (1980) 217–225.
[23] F. Soltani, D. Toghraie, A. Karimipour, Experimental measurements of thermal [52] J. Thaker, J. Banerjee, On intermittent flow characteristics of gas–liquid two-phase
conductivity of engine oil-based hybrid and mono nanofluids with tungsten oxide flow, Nucl. Eng. Des. 310 (2016) 363–377.
(WO3) and MWCNTs inclusions, Powder Technol. 371 (2020) 37–44. [53] A. Arabi, Y. Salhi, E.K. Si-Ahmed, J. Legrand, Influence of a sudden expansion on
[24] G. Sriharan, S. Harikrishnan, H.M. Ali, Experimental investigation on the effec- slug flow characteristics in a horizontal two-phase flow: a pressure drop fluctua-
tiveness of MHTHS using different metal oxide-based nanofluids, J. Therm. Anal. tions analysis, Meccanica 53 (13) (2018) 3321–3338.
Calorim. (2020), https://doi.org/10.1007/s10973-020-09779-5 (in press). [54] A.H. Zitouni, A. Arabi, Y. Salhi, Y. Zenati, E.K. Si-Ahmed, J. Legrand, Slug length
[25] H.A. Tariq, M. Anwar, A. Malik, H.M. Ali, Hydro-thermal performance of normal- and frequency upstream a sudden expansion in gas-liquid intermittent flow,
channel facile heat sink using TiO, J. Therm. Anal. Calorim. (2020), https://doi. Experimental and Computational Multiphase flow 3 (2) (2021) 124–130.
org/10.1007/s10973-020-09838-x (in press). [55] Z. Ruder, T.J. Hanratty, A definition of gas-liquid plug flow in horizontal pipes,
[26] A. Chen, T.F. Lin, H.M. Ali, W.M. Yan, Experimental study on bubble character- Int. J. Multiphase Flow 16 (2) (1990) 233–242.
istics of time periodic subcooled flow boiling in annular ducts due to wall heat flux [56] J.F. Netto, J. Fabre, L. Peresson, Shape of long bubbles in horizontal slug flow, Int.
oscillation, Int. J. Heat Mass Transf. 157 (2020) 119974. J. Multiphase Flow 25 (6–7) (1999) 1129–1160.
[27] H. Ali, Hybrid Nanofluids for Convection Heat Transfer, 1st Ed, Academic Press, [57] M.G. Conte, G.A. Hegde, M.J. da Silva, A.K. Sum, R.E. Morales, Characterization of
Elsevier, 2020, pp. 1–300, https://doi.org/10.4018/978-1-7998-1546-4. slug initiation for horizontal air-water two-phase flow, Exp. Thermal Fluid Sci. 87

11
A. Arabi, et al. International Communications in Heat and Mass Transfer 118 (2020) 104841

(2017) 80–92. slug frequency correlations for two-phase gas-liquid horizontal slug flow, ASME
[58] F. Humami, O. Dinaryanto, A.Z. Hudaya, A. Widyatama, Indarto, Deendarlianto, 2014 4th Joint US-European Fluids Engineering Division Summer Meeting collo-
Experimental study on the characteristics of flow pattern transitions of air-water cated with the ASME 2014 12th International Conference on Nanochannels,
two-phase flow in a horizontal pipe, AIP Conference Proceedings, 2001 AIP Microchannels, and Minichannels, American Society of Mechanical Engineers,
Publishing, 2018, August, p. 030005 1. 2014, Augustpp. V002T06A009-V002T06A009.
[59] J. Thaker, J. Banerjee, Transition of plug to slug flow and associated fluid dy- [86] J. Nair, E. Pereyra, C. Sarica, A statistical evaluation of slug frequency models for
namics, Int. J. Multiphase Flow 91 (2017) 63–75. air-water flow in 6-in. ID near horizontal pipes, 10th North American Conference
[60] A. Arabi, Y. Salhi, Y. Zenati, E.K. Si-Ahmed, J. Legrand, On gas-liquid intermittent on Multiphase Technology, BHR group, 2016, December.
flow in a horizontal pipe: influence of sub-regime on slug frequency, Chem. Eng. [87] R. Kong, A. Rau, S. Kim, S. Bajorek, K. Tien, C. Hoxie, Experimental study of
Sci. 211 (2020) 115251. horizontal air-water plug-to-slug transition flow in different pipe sizes, Int. J. Heat
[61] D. Barnea, N. Brauner, Holdup of the liquid slug in two phase intermittent flow, Mass Transf. 123 (2018) 1005–1020.
Int. J. Multiphase Flow 11 (1) (1985) 43–49. [88] J. Thaker, J. Banerjee, Influence of intermittent flow sub-patterns on erosion-
[62] Y. Taitel, A.E. Dukler, A model for predicting flow regime transitions in horizontal corrosion in horizontal pipe, J. Pet. Sci. Eng. 145 (2016) 298–320.
and near horizontal gas-liquid flow, AICHE J. 22 (1) (1976) 47–55. [89] M.G. Hubbard, An Analysis of Horizontal Gas-Liquid Slug Flow, Doctoral dis-
[63] V. Bertola, Experimental characterization of gas–liquid intermittent subregimes by sertation, University of Houston, 1965.
phase density function measurements, Exp. Fluids 34 (1) (2003) 122–129. [90] N.I. Heywood, J.F. Richardson, Slug flow of air—water mixtures in a horizontal
[64] S.Q. Wang, K.W. Xu, H.B. Kim, Slug flow identification using ultrasound Doppler pipe: determination of liquid holdup by γ-ray absorption, Chem. Eng. Sci. 34 (1)
velocimetry, Int. J. Heat Mass Transf. (2019), https://doi.org/10.1016/j. (1979) 17–30.
ijheatmasstransfer.2019.119004 119004. (article in press). [91] O.J. Nydal, S. Pintus, P. Andreussi, Statistical characterization of slug flow in
[65] O. Dinaryanto, Y.A.K. Prayitno, A.I. Majid, A.Z. Hudaya, Y.A. Nusirwan, horizontal pipes, Int. J. Multiphase Flow 18 (3) (1992) 439–453.
A. Widyaparaga, Experimental investigation on the initiation and flow develop- [92] I.G. Manolis, M.A. Mendes-Tatsis, G.F. Hewitt, The effect of pressure on slug fre-
ment of gas-liquid slug two-phase flow in a horizontal pipe, Exp. Thermal Fluid quency in two-phase horizontal flow, Multiphase Flow 1995, 1995, pp. 347–354.
Sci. 81 (2017) 93–108. [93] F.S. Silva, V.M. Toledo, D.P. Quinto, J.C. Maya, Experimental slug flow char-
[66] A.O. Mohmmed, H.H. Al-Kayiem, M.S. Nasif, R.W. Time, Effect of slug flow fre- acterization in a horizontal pipe, Exp. Heat Transf. Fluid Mech. Thermodyn.
quency on the mechanical stress behavior of pipelines, Int. J. Press. Vessel. Pip. (1997) 893–900.
172 (2019) 1–9. [94] M. Fossa, G. Guglielmini, A. Marchitto, Intermittent flow parameters from void
[67] Y. Salhi, Contribution théorique et expérimentale à l'étude des phénomènes de fraction analysis, Flow Meas. Instrum. 14 (4–5) (2003) 161–168.
transition d'un écoulement stratifié à l'écoulement poche/bouchon dans une con- [95] B.J. Azzopardi, Drops in annular two-phase flow, Int. J. Multiphase Flow 23 (7)
duite horizontale en présence de singularité. PhD thesis, USTHB, Algiers. Algeria, (1997) 1–53.
2010. [96] E. Wren, G. Baker, B.J. Azzopardi, R. Jones, Slug flow in small diameter pipes and
[68] E. Al-Safran, B. Gokcal, C. Sarica, High viscosity liquid effect on two-phase slug T-junctions, Exp. Thermal Fluid Sci. 29 (8) (2005) 893–899.
length in horizontal pipes, In 15th International Conference on Multiphase [97] B. Gokcal, A. Al-Sarkhi, C. Sarica, E.M. Al-safran, Prediction of slug frequency for
Production Technology, BHR Group, 2011, June. high-viscosity oils in horizontal pipes, SPE Projects, Facilities & Construction 5
[69] V. Bertola, Slug velocity profiles in horizontal gas–liquid flow, Exp. Fluids 32 (6) (03) (2010) 136–144.
(2002) 722–727. [98] A. Azzi, A. Al-Attiyah, L. Qi, W. Cheema, B.J. Azzopardi, Gas–liquid two-phase
[70] H.H. Al-Kayiem, A.O. Mohmmed, Z.I. Al-Hashimy, R.W. Time, Statistical assess- flow division at a micro-T-junction, Chem. Eng. Sci. 65 (13) (2010) 3986–3993.
ment of experimental observation on the slug body length and slug translational [99] R.W. Lockhart, R.C. Martinelli, Proposed correlation of data for isothermal two-
velocity in a horizontal pipe, Int. J. Heat Mass Transf. 105 (2017) 252–260. phase, two-component flow in pipes, Chem. Eng. Prog. 45 (1) (1949) 39–48.
[71] K.H. Bendiksen, M. Langsholt, L. Liu, An experimental investigation of the motion [100] R. Schulkes, Slug frequencies revisited, In 15th International Conference on
of long bubbles in high viscosity slug flow in horizontal pipes, Int. J. Multiphase Multiphase Production Technology, BHR Group, 2011, June.
Flow 104 (2018) 60–73. [101] F.E. Vicencio, F.A. Schneider, C. Cozin, F.A. Barbuto, M.J. Da Silva, R.E. Morales,
[72] G.H. Abdul-Majeed, A.M. Al-Mashat, A unified correlation for predicting slug li- An experimental characterization of horizontal gas-liquid slug flow, ASME 2015
quid holdup in viscous two-phase flow for pipe inclination from horizontal to International Mechanical Engineering Congress and Exposition, American Society
vertical, SN Appl. Sci. 1 (1) (2019) 71. of Mechanical Engineers, 2015, November(pp. V07BT09A020-V07BT09A020).
[73] G.H. Abdul-Majeed, M. Firouzi, The suitability of the dimensionless terms used in [102] E.M. Al-Safran, Probabilistic modeling of slug frequency in gas/liquid pipe flow
correlating slug liquid holdup with flow parameters in viscous two-phase flows, using the Poisson probability theory, J. Pet. Sci. Eng. 138 (2016) 88–96.
Int. Commun. Heat Mass Transf. 108 (2019) 104323. [103] M. Abdulkadir, V. Hernandez-Perez, I.S. Lowndes, B.J. Azzopardi, E. Sam-
[74] E.J. Greskovich, A.L. Shrier, Slug frequency in horizontal gas-liquid slug flow, Mbomah, Experimental study of the hydrodynamic behaviour of slug flow in a
Indust. Eng. Chem. Process Des. Dev. 11 (2) (1972) 317–318. horizontal pipe, Chem. Eng. Sci. 156 (2016) 147–161.
[75] S. Mohammadi, M. Papa, E. Pereyra, C. Sarica, Genetic algorithm to select a set of [104] S. Farsetti, S. Farisè, P. Poesio, Experimental investigation of high viscosity oil–air
closure relationships in multiphase flow models, J. Pet. Sci. Eng. 181 (2019) intermittent flow, Exp. Thermal Fluid Sci. 57 (2014) 285–292.
106224. [105] Y.D. Baba, A.E. Archibong, A.M. Aliyu, A.I. Ameen, Slug frequency in high visc-
[76] C. Dong, L. Lu, X. Wang, Experimental investigation on non-boiling heat transfer osity oil-gas two-phase flow: experiment and prediction, Flow Meas. Instrum. 54
of two-component air-oil and air-water slug flow in horizontal pipes, Int. J. (2017) 109–123.
Multiphase Flow 119 (2019) 28–41. [106] S.L. Kokal, An experimental study of two phase flow in inclined pipes. PhD thesis,
[77] G.H. Carvalho, L.H. Melo Lima, A Review of Frequency Correlations for the University of Calgary, Alberta. Canada., 1987.
Intermittent Gas-Liquid Flow in Horizontal Pipes. 17th Brasilian Congress of [107] V. Bertola, E. Cafaro, Slug frequency measurement techniques in horizontal gas-
Thermal Sciences and Engineering. November 25th–28th, 2018, Aguas de Lindoia, liquid flow, AIAA J. 40 (5) (2002) 1010–1012.
Brasil, 2018. [108] M. Abdulkadir, V. Hernandez-Perez, C.A. Kwatia, B.J. Azzopardi, Interrogating
[78] Y. Zhao, H. Yeung, L. Lao, Slug frequency in high viscosity liquid and gas flow in flow development and phase distribution in vertical and horizontal pipes using
horizontal pipes, In 16th International Conference on Multiphase Production advanced instrumentation, Chem. Eng. Sci. 186 (2018) 152–167.
Technology, BHR Group, 2013, June. [109] M. Abdulkadir, D.G. Jatto, L.A. Abdulkareem, D. Zhao, Pressure drop, void frac-
[79] A. Archibong-Eso, Y. Baba, A. Aliyu, Y. Zhao, W. Yan, H. Yeung, On slug frequency tion and flow pattern of vertical air–silicone oil flows using differential pressure
in concurrent high viscosity liquid and gas flow, J. Pet. Sci. Eng. 163 (2018) transducer and advanced instrumentation, Chem. Eng. Res. Des. 159 (2020)
600–610. 262–277.
[80] G.J. Zabaras, Prediction of slug frequency for gas/liquid flows, SPE J. 5 (03) [110] J.M. Mandhane, G.A. Gregory, K. Aziz, A flow pattern map for gas—liquid flow in
(2000) 252–258. horizontal pipes, Int. J. Multiphase Flow 1 (4) (1974) 537–553.
[81] V. Hernandez Perez, Gas-Liquid Two-Phase Flow in Inclined Pipes, Doctoral dis- [111] N. Andritsos, L. Williams, T.J. Hanratty, Effect of liquid viscosity on the stratified-
sertation, University of Nottingham, 2008. slug transition in horizontal pipe flow, Int. J. Multiphase Flow 15 (6) (1989)
[82] G.A. Gregory, D.S. Scott, Correlation of liquid slug velocity and frequency in 877–892.
horizontal cocurrent gas-liquid slug flow, AICHE J. 15 (6) (1969) 933–935. [112] M.J. Vaze, J. Banerjee, Prediction of liquid height for onset of slug flow, Can. J.
[83] X. Wang, L. Guo, X. Zhang, An experimental study of the statistical parameters of Chem. Eng. 90 (5) (2012) 1295–1303.
gas–liquid two-phase slug flow in horizontal pipeline, Int. J. Heat Mass Transf. 50 [113] B.D. Woods, Z. Fan, T.J. Hanratty, Frequency and development of slugs in a
(11−12) (2007) 2439–2443. horizontal pipe at large liquid flows, Int. J. Multiphase Flow 32 (8) (2006)
[84] R.H. Shea, H. Eidsmoen, M. Nordsveen, J. Rasmussen, Z. Xu, J. Nossen, Slug fre- 902–925.
quency prediction method comparison, proceedings of the 4th north American [114] M. Lin, Y. Liu, Y. Hu, D. Che, Influence of the gas and liquid superficial velocity on
conference on multiphase technology, 2004, June, pp. 227–237. slug frequency, AIP Conference Proceedings, 1547 AIP, 2013, July, pp. 253–263 1.
[85] M.M. Antunes, C. Cozin, F.A. Barbuto, R.E. Morales, H.T. Rodrigues, Analysis of

12