Author’s Accepted Manuscript

Influence of intermittent flow sub-patterns on

erosion-corrosion in horizontal pipe

Jignesh Thaker, Jyotirmay Banerjee

www.elsevier.com/locate/petrol

PII: S0920-4105(16)30169-3
DOI: http://dx.doi.org/10.1016/j.petrol.2016.05.006
Reference: PETROL3449
To appear in: Journal of Petroleum Science and Engineering
Received date: 30 December 2015
Revised date: 12 April 2016
Accepted date: 6 May 2016
Cite this article as: Jignesh Thaker and Jyotirmay Banerjee, Influence of
intermittent flow sub-patterns on erosion-corrosion in horizontal pipe, Journal of
Petroleum Science and Engineering,
http://dx.doi.org/10.1016/j.petrol.2016.05.006
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Influence of intermittent flow sub-patterns on erosion-corrosion in
horizontal pipe1
Jignesh Thaker, Jyotirmay Banerjee*

Department of Mechanical Engineering, Sardar Vallabhbhai National Institute of Technology (SVNIT), Surat-
395007, Gujarat, India
*
Corresponding author. Tel.: +91 261 2204145; fax:+91 261 2228394; jbaner@med.svnit.ac.in

Abstract

Thinning of pipe wall due to erosion and corrosion has resulted in many catastrophic failures and huge
economic losses in petroleum refineries (Wood et al., 2013) and nuclear power stations (Ahmed, 2012). The
intermittent structure of two-phase flow through pipe (intermittent flow regime) is an important factor
responsible for such erosion and corrosion. The present research is an effort to develop a regime map showing
possible erosion and corrosion phenomena due to intermittent flows in pipe. In this direction, flow visualization
experiments are carried out for three distinct sub-regimes of intermittent flow namely, plug flow with or without
bubble rim, less aerated slug flow and highly aerated slug flow. Images captured at a rate of 1600 frames per
second at 288 diameter downstream of inlet are utilised for analysing the dynamics of intermittent flow structure
in terms of expansion, contraction, breaking, coalescence, collapse, and collision of bubbles. Quantitative
measurements of intermittent flow characteristics (including plug/slug frequency, plug/slug velocity and length
of liquid plug/slug) and their influence on erosion-corrosion phenomena in the pipe are reported in detail. Based
on these qualitative and quantitative analysis, a regime map for prediction of erosion-corrosion phenomena in
pipes is established as a function of inlet flow conditions for both the phases. Transition boundaries for four
distinct erosion-corrosion phenomena: shear stress induced erosion, cavitation erosion, liquid impact induced
erosion and flow accelerated corrosion are represented in the map and eight distinct regimes of erosion-
corrosion are illustrated. The regime map is represented in terms of non-dimensional superficial Reynolds
numbers of both the phases to account for pipe diameter, flow rate and fluid viscosity. This erosion-corrosion
regime map developed in this research will immensely aid to the effective design of piping systems and
optimization of operating conditions for safer operation of petroleum refineries and nuclear power stations.

Keywords

Intermittent flow; flow visualization; flow characteristics; erosion-corrosion regime map.

1. Introduction

Material degradation or wall thinning due to erosion-corrosion in piping system represents

one of the major problems in petroleum industries, nuclear power plants, chemical industries,
geothermal industries and desalination plants. Such erosion-corrosion normally occurs on
various piping components including tees, elbows, downstream of control valves, flow
elements, reducers or orifices. Catastrophic failures with serious fatalities in past due to
erosion-corrosion in piping systems have been reported by Wood et al. (2013) and Ahmed
(2012). Preliminary evaluation of these failures indicated that flow accelerated corrosion
(FAC) and liquid impact induced erosion (LIIE) were the likely failure mechanisms. Such
mechanisms have been related in literature to the intermittent structure of two-phase pipe

1
Project sponsored by Science and Engineering Research Board (SERB), Government of India.
flow (Jepson, 1989; Sun and Jepson, 1992; Bozzini et al., 2003; Shawki, 2011; Ahmed, 2012;
and Wood et al., 2013). The pertinent question that arises in this regard is how the
intermittent structure of two-phase flow induces flow acceleration and liquid impact on pipe.
In order to address this question, it is required to understand the different sub-patterns of
intermittent flow structure in terms of their dynamics (which includes dynamics of aeration
and bubble shape geometry) and their influence on periodic acceleration of flow and impact
of liquid on piping systems.
Extensive studies are available in literature on estimating the corrosion rate (especially
due to FAC) as a function of local flow and mass transfer parameters without going into the
details on the dynamics of flow structure. Poulson, 1990; Shirazi et al., 1995; Chexal et al.,
1996; Kuo-Tong et al., 1998; Poulson, 1999; Ferng et al., 2000; Yuh et al., 2008; Li et al.,
2009; Yu et al., 2009 and Ahmed et al., 2014 conducted experimental studies for evaluating
the correct FAC wear rate and developed appropriate model for specific industrial
applications. More recently, Ahmed et al. (2014) reported that the pipe wall thinning rate
(due to FAC) mainly depends on complex interaction of several parameters such as material
composition, water chemistry, and hydrodynamics. Significant research were also reported on
investigating the effect of fluid chemical properties on FAC. However, research on
hydrodynamic influence of intermittent type two-phase flow on FAC is rarely available. The
dynamic interaction between gas phase and liquid turbulence structures associated with such
flows aggravate the corrosion problem (Kim et al., 2007).
Keating and Nesic, 1999; Bozzini et al., 2003; Chang et al., 2006; Ferng, 2008 have
predicted the mass transfer and corrosion rate due to FAC using computational fluid dynamic
(CFD) analysis. In formulating the CFD codes, consideration is made to the hydrodynamic
parameters affecting the mass transfer rate of the corrosion products to the bulk fluid and
consequently the FAC rate. These hydrodynamic parameters include the flow structure,
phasic velocity, pipe roughness, piping geometry and void fraction (Ahmed et al., 2014).
Such hydrodynamic effects of the working fluid on FAC have been investigated by a few
researchers (Bozzini et al., 2003; Chang et al., 2006; Ferng, 2008) using CFD. More recently,
Ferng (2008) developed an approach that used an erosion/corrosion model and three-
dimensional single and two-phase flow models to predict location of FAC in piping systems.
They identified that accurate modelling of secondary vortices (secondary flows) in single
phase flow and flow structures in two-phase flow as the two main challenges that need to be
addressed before dealing with FAC in piping systems. Due to lack of information in literature
on the dynamics of interfacial structure of intermittent type two-phase flow and its influence
on FAC and LIIE, accurate models for erosion-corrosion in two-phase pipe flow are yet to be
established (Ferng, 2008; Prabha et al., 2014; Ahmed et al., 2014).
On other hand, Taitel and Dukler, 1976; Fabre and Liné, 1992; De Henau and Raithby,
1995; Kadri et al., 2009; and Kjeldby et al., 2013 have presented simplified theoretical
models for slug type intermittent flow. Complete mathematical treatment of intermittent flow
is difficult due to the complexity of interaction between dispersed and continuous phases and
among the dispersed phases. The dispersed phase can be characterized by two internal length
scales and velocity scales: one due to small bubbles in liquid slugs/plugs and another
associated with the large bubbles between two liquid slugs. Empirical observations and
correlations for main characteristics of such flows are extremely important to close the
models and to validate the numerical results. The most significant parameters, which
characterize intermittent flow, are the gas and liquid phase distribution (void fraction and
liquid holdup), liquid plug/slug velocity and its fluctuations, liquid plug/slug frequency,
liquid plug/slug length, and turbulent characteristics of mass, momentum and energy transfer
at the interface (Sharma et al., 1998). These parameters are transient due to the intermittent
and irregular nature of flow. The knowledge of time-averaged values of these parameters is
not always sufficient for design, and statistical information might be relevant (Carneiro et al.,
2011). Experimental research towards measurement of slug flow parameters in horizontal
macro scale pipes using different measurement techniques reported in literature is
summarised in Table 1 according to author/year, fluids used, test rig diameter and major
research contribution. Most of the experiments reported in literature (Table 1) were
conducted on pipe diameters ≤ 50 mm and on transparent (plexiglas or acylic) pipes for better
visualisation of slug flow dynamics.
There is however, no literature available that attempts to correlate the dynamics of
intermittent flow structures to the possible erosion-corrosion phenomena in pipe flow.
Moreover, intermittent regime of two-phase flow includes both slug and plug flow patterns. It
is thus required to establish flow characteristics for the entire regime of intermittent flow
including the plug flow. In addition, no work is reported in literature that correlates the
intermittent flow parameters with erosion-corrosion phenomena in piping systems.
Present work is an attempt to bridge this research gap by analysing the dynamic changes
in interfacial structure of intermittent flow regime and correlating them to the possible
erosion-corrosion phenomena in piping system. In this direction, two-phase air-water flow
visualisation experiments are conducted on 25 mm inner diameter horizontal transparent pipe
of 14 m length. The interfacial structure of intermittent flow in terms of break up, expansion,
contraction, coalescence, collapse and collision of bubbles is illustrated. The possible
corrosion-erosion phenomena inside the piping systems are then predicted based on these
interfacial mechanisms of the flow. The present work also deals with the analysis of flow
characteristics in three distinct sub-patterns of intermittent flow (namely, plug flow with or
without bubble rim, less aerated slug flow and highly aerated slug flow). The influence of
these flow characteristics are then correlated to possible erosion-corrosion phenomena in
piping systems. The correlations are established in the present work in terms of non-
dimensional superficial Reynolds numbers to account for pipe diameter, flow rate and fluid
viscosity. It is worth mentioning here that the physics of slug flow transition, dynamics of
slug flow and variation of slug flow parameters (slug frequency, slug velocity and slug
length) discussed in literature (see Table 1) on different pipe diameters and different fluids
show similar trend. Hence, the correlation between intermittent flow dynamics and erosion-
corrosion established in the present work on 25 mm internal diameter pipe can be used for the
realistic pipe dimensions of petroleum and nuclear industries.
In what follows this paper, details of the experimental test facility is described in section
2. Experimental flow conditions used for further analysis and associated measurement
uncertainties are reported in Section 3. Interfacial dynamics for three sub-regimes of
intermittent flow are analysed using visual images and are correlated to erosion-corrosion
phenomena in section 4. Measurement procedure used for evaluating the intermittent flow
characteristics are reported in Section 5. Detailed analysis of intermittent flow characteristics
in terms of inlet flow conditions and their influence on erosion-corrosion phenomena in
piping systems are reported in Section 6. Based on the extensive database of the present
research, a regime map is established in Section 7 for the prediction of erosion-corrosion in
terms of inlet flow conditions (ReSL and ReSG). Major observations of the present work are
highlighted in Section 8.

2. Experimental flow facility for two-phase flow measurements

The two-phase experimental flow facility illustrated in Fig. 1 mainly includes two-phase flow
test rig (TPFTR) and high speed photography system (HSPS).

2.1 Two-phase flow test rig (TPFTR)

Schematic diagram of the two-phase air-water flow test rig is shown in Fig. 2. The test rig
(TPFTR) contains the following major components:
 Air flow loop,
 Water flow loop,
 SCADA (Supervisory Control and Data Acquisition),
 Two-phase flow mixture and
 Test section for two-phase flow measurements.

2.1.1 Air flow loop

Dry air is supplied by the screw compressor and is stored in receiver tank. This dry air from
the receiver tank passes through the air filter, pressure regulator, Coriolis mass flow meter,
two-phase mixture, and finally enters into the test section (shown in Fig. 2). Air is filtered
and condensates are removed using a coalescing filter. Air pressure is controlled by the
pressure regulator, while the mass flow rate of air is measured by a Coriolis flow meter with
an accuracy of ±0.35%. Coriolis meter is connected to a Field-Mount Transmitter that
conditions the flow information for the data acquisition system. Air is delivered to the two-
phase mixture by flexible tubing.

2.1.2 Water flow loop

Distilled water pumped by the Centrifugal pump passes through the filter (strainer), water
bath heat exchanger and Coriolis flow meter, connected to a digital Field-Mount Transmitter.
A water bath heat exchanger is used to remove the pump heat and the heat added during the
experiment to maintain a constant inlet water temperature. Water mass flow rate is measured
using Coriolis flow meter with an accuracy of ±0.1%.

2.1.3 Supervisory control and data acquisition system (SCADA)

Mass flow of air and water are regulated by the pneumatic control valves which are
monitored by Proportional Integral Derivative (PID) controllers using Supervisory Control
and Data Acquisition (SCADA) system through workstation. Flowchart of this process is
shown in Fig. 3.
2.1.4 Two-phase flow mixture

In order to attain proper mixing of two phases, a two-phase flow mixture is placed at the
entry of the test section (see Fig. 4 (a)). Detailed dimensions of two-phase flow mixture is
shown in Fig. 4 (b). The mixture consists of a perforated stainless steel tube inserted into the
liquid stream. On the periphery of the tube, four holes are drilled at 90° interval and this
pattern is repeated at fifteen equally spaced locations along the length of the stainless steel
tube as shown in Fig. 4. Air flows through this tube and enters the test section through these
small holes on the periphery of the tube and finally mixes with the surrounding water. The
tube is silver-soldered at the end for preventing the flow of air.

2.1.5 Test section for two-phase flow measurements

Test section contains three 14 m long parallel transparent acrylic pipes. Internal diameters of
these pipes are 50 ±0.20 mm, 25 ±0.15 mm and 12 ±0.10 mm each. Two-phase flow
experiments reported here are performed on pipe of 25 ±0.15 mm inner diameter. These pipes
are laid on 15 m long steel frame structure and are supported by nine bakelite made stationary
feet. The entire frame structure rests on a stationary foot as shown in Fig. 2. The stationary
foot has a pivot for the inclination of the entire test setup using chain pulley. Present
experiments are conducted for horizontal configuration. Such horizontal orientation is
maintained using spirit level and flexible tube water level techniques. Mixture of air and
water passes through the Mixing Zone, Developing Zone and Visualization Zone.
Visualization zone is located at 6 m from the inlet which allows sufficient length for the plugs
or slugs to form and grow. Two-phase air-water mixture coming out of the test section enters
into a two-phase separation tank shown in Fig. 2. Air from the mixture is vented to the
surrounding and the water is collected in the receiver for recirculation. Temperature sensor
(PT-100) is placed at the inlet of the test section to monitor the constant water temperature
maintained by the water bath heat exchanger.

2.2 High Speed Photography System (HSPS)

The high speed photography system (shown in Fig. 5) consists of a Photron made FASTCAM
camera (speed: 30 to 100000 FPS, resolution: 1280×1280) with tripod, workstation, portable
white board and a scale.
The camera is mounted on the tripod stand to rotate the camera in any direction and to adjust
the height of the camera from ground level as shown in Fig. 5. The tripod is facilitated with
spirit level to align the camera and tripod stand in perfect horizontal position. Complete
control of the camera operating system (including frame rate triggering, adjustment image
resolutions and shutter speed) is facilitated through the workstation using Photron
FASTCAM Viewer (windows based software). A 1.6 m long horizontal scale (supported by
two portable stands) is used (as shown in Fig. 5) for quantitative estimation of intermittent
flow characteristics from the images.

3. Experimental flow conditions and measurement uncertainties

Experimental flow visualisation is carried out for different inflow conditions of both the
phases represented in terms of superficial Reynolds number of gas (ReSG) and liquid (ReSL).
The range of ReSG and ReSL for intermittent flow regime is carefully selected based on the
generalized two-phase flow regime map of Ghazar and Tang (2007) and Vaze and Banerjee
(2011). These maps developed in terms of ReSG and ReSL show the transition boundaries for
two-phase flow patterns in horizontal pipe. In these flow regime maps, the classification of
intermittent flow is shown as plug flow and slug flow based on the presence of aeration inside
the liquid pockets as shown in Fig. 6.
Investigations on classification of slug flow (sub-pattern of intermittent flow) was reported by
Thaker and Banerjee (2015) based on the morphology of air-water interface and the presence
of aeration inside the liquid slug. Five distinct sub-regimes of slug flow were classified,
namely, slug formation zone; plug and less aerated slug zone; less aerated slug zone; highly
aerated slug zone; and highly aerated slug with wavy interface zone. In the present work, the
interfacial dynamics are investigated for plug flow (with or without bubble rim), less aerated
slug flow and highly aerated slug flow (shown in Fig. 6) in order to correlate to the possible
erosion-corrosion phenomena in pipes. The visualised images of liquid plug/slug front
(bubble tail) and liquid plug/slug tail (bubble nose) shown in Fig. 6. Fig. 6 (a) and (c) depicts
plug flow with or without bubble rim (discuss later in section 4.1). Plug and slug flow are
distinguishable by the presence of aeration inside the liquid slug front (shown in Fig. 6 (a),
(c), (e) and (g)). Less Aerated Slug (LAS) and Highly Aerated Slug (HAS) is differentiable
based on the presence of eddies (swirls) associated with bubbles inside the liquid slugs. LAS
flow (transition from plug to slug flow) contains less aeration (absence of eddies) and distinct
bubbles in the liquid slugs (shown in Fig. 6 (e) and (f)) while HAS flow contains larger
aeration and large number of indistinguishable bubbles with eddies (encircled in Fig. 6 (g)) in
liquid slug front.
The transition lines between plug, LAS and HAS sub-regimes are shown on the flow regime
map reported by Thaker and Banerjee (2015) in Fig. 7. Transition boundary between
stratified or wavy-stratified flow to less aerated slug flow and stratified or wavy-stratified
flow to highly aerated slug flow are depicted by lines AB and BA' respectively in Fig. 7. Line
CC' in Fig. 7 represents the transition boundary between plug flow and less aerated slug flow.
Less aerated slug flow regime is separated from highly aerated slug flow regime by thick
transition line BB' in Fig. 7. Visual images for each sub-patterns are also shown in Fig. 7. As
a part of validation for the present experimental analysis, transition line between stratified
flow and slug flow or the condition for onset of slugging (AA' shown in Fig. 7) is compared
with the results of Lin (1985) and Andritsos et al. (1989). For comparison purpose, the
present experimental results are represented in terms of superficial velocities of both the
phases (calculated from the inlet mass flow condition) in Fig. 8.
The condition for onset of slugging obtained from the present experiments shown in Fig. 8
are in close agreement with those reported by Lin (1985). For lower values of superficial gas
velocity, experimental results (shown in Fig. 8) agree closely with those reported by
Andritsos et al. (1989). For higher superficial gas velocity, the experimental results of
Andritsos et al. (1989) show a larger liquid flow rate required for initiation of slug as
compared to the present experiments and those reported by Lin (1985).

A set of 54 combinations of inflow conditions are considered (shown in Fig. 9) for examining
the interfacial dynamics associated with the three sub-patterns (plug, LAS and HAS) of
intermittent flow and correlating them with possible erosion-corrosion phenomena in piping
system. These 54 data points (tabulated in Table 2) are chosen to cover the entire range of
Reynolds numbers corresponding to intermittent flow regime shown in Fig. 9.
Uncertainty of flow measurement in terms of superficial Reynolds number of liquid (ReSL)
and gas (ReSG) are calculated following the procedure described in UKAS (2007) and are
reported in Table 3.
4. Analysis of internal structure of intermittent flow sub-patterns
The failures in major oil line or steam line due to erosion-corrosion are undesirable for
petroleum and nuclear industries. It is important to quantify the influence of intermittent flow
on erosion-corrosion under varying conditions so that effective corrosion control can be
achieved. In this direction, Kaul (1995) made an effort to develop corrosion inhibitors for oil-
gas pipelines and she found that the corrosion inhibitors do not work well for intermittent
flow conditions due to their uneven flow structures. The first step in determining the
predominant corrosion mechanisms is to understand the dynamics of intermittent flow. This
section deals with the flow dynamics of three distinct intermittent flow sub-patterns and their
influence on possible erosion-corrosion phenomena in piping systems.

4.1 Plug flow with or without bubble rim

In this sub-regime, the liquid phase/pocket (called as liquid plug) is observed free of bubbles
and occupies the entire cross-sectional area of the pipe. The liquid plugs are separated by
elongated or shorter bubbles (see Fig. 6). Plug flow patterns are observed at very low values
of ReSG (see Fig. 7). Internal flow structure of this sub-regime is discussed based on analysis
of images captured for fourteen different combinations of ReSL and ReSG as shown in Fig. 10.
Since the length of plug flow exceeds the visualization or observation window of the camera,
images are presented in Fig. 10 for two essential portions of plug flow where major changes
in shape of the bubble or liquid plug are observed. These are: near the bubble tail (liquid plug
front) and near the bubble nose (liquid plug tail).
The shape of bubble geometry (at bubble tail and at bubble nose) for two constant values of
ReSG (for varying ReSL) is shown in Fig. 10 (a)-(g) and (a')-(g'). It is observed that the nose of
the plug bubble occupies more cross-sectional area and becomes thinner towards the bubble
tail. This is due to higher acceleration of liquid plug (getting inertia from the air) near the
bubble tail which tries to overcome the surface tension force between plug bubble and pipe
wall. Due to buoyancy continuous liquid film is always present at the bottom of the pipe in
plug flow and the plug bubble always remain present near the upper periphery of the pipe
wall. Upper wall of the pipe is periodically influenced by flowing liquid and gas phases at
different phasic velocities. This periodicity over long time period can lead to the onset of the
erosion-corrosion process by damaging the protective films on internal surface of the upper
pipe wall.
The length of the bubble tail gets reduced with increase in ReSL as shown in Fig. 10 (a)-
(g) and (a')-(g'). This occurs due to the augmentation in turbulent mixing inside the liquid
plug with increase in ReSL. The turbulent mixing counteracts the buoyancy and surface
tension effect between the plug bubble and the pipe wall. It is also observed that the length of
liquid plug and the elongated plug bubble gradually decreases with increase in ReSL and there
is an increase in rate of formation of plug. This rapid rate of plug formation is responsible for
large deviation in gas-liquid flow rates and may lead to Shear Stress Induced Erosion on the
internal surface of the pipe. On other hand, with increase in ReSG (compare Fig. 10 (a)-(g)
with (a')-(g')) a reduction in length of bubble tail is observed due to higher momentum of air
which overcomes the surface tension effect. Further increase in ReSG (beyond 1147) leads to
entrainment of bubbles inside the liquid plug front. This is a transition phase where plug flow
begins to change into less aerated slug flow (discussed in sub-section 4.2).
One major observation is that the elongated plug bubble tail continuously changes size
along the length of the pipe for each combination of ReSL and ReSG. Plug bubble tail
(especially bubble rim) has a tendency to increase and decrease in length, sometimes this
bubble rim breaks into small bubbles. These phenomena are clearly demonstrated through the
three videos uploaded with this paper.
These videos are recorded a for flow condition of ReSG = 765 and ReSL = 9069. The growth
(elongation) of the bubble rim is due to the dominance of surface tension between the bubble
and the upper periphery of the pipe wall and results in retardation of motion of the bubble at
the bubble tail (because of the shearing of bubble at the upper periphery of the pipe wall).
The bubble rim is stretched until the sharp end is reached. Once the sharp edge of bubble rim
forms or once the stretching of this bubble rim is sufficient to overcome the surface tension
force (due to the momentum of the following liquid plug), the bubble rim reduction process
starts. However, as the flow progresses, the length of bubble rim gets reduced gradually due
to the dominance of upstream liquid plug momentum over surface tension between the
bubble rim and the pipe wall. When this bubble rim completely diminishes, a long bubble tail
with the sharp bubble rim is formed again. Due to such asymmetric expansion and
contraction of plug bubble rim, modelling of plug flow pattern is difficult and such
continuous expansion and contraction of plug bubble rim at upper wall of the pipe leads to
enhancement of erosion-corrosion of piping material.
Sometime, due to the relative motion between the gas and liquid phase and associated
interfacial shear, the interface of the long bubble rim tends to be wavy (observed in videos).
When this wave height increases beyond the local plug bubble depth, small gas pocket or
bubble is detached from the bubble rim. Separation of bubble from the bubble rim are also
observed in Figs. 10 (d), (b')-(g'). Separated bubble travels to the upper periphery of the pipe
inside liquid plug due to buoyancy. Frequent drying and wetting of pipe surface due to such
motion of bubble in liquid slug for long period of time can cause corrosion of the internal
pipe surface. Moreover, pressure inside such bubble is higher compared to the surrounding
liquid pressure creating resistance to the motion of bubble. This leads to retardation of the
bubble motion compared to liquid plug. This retardation results in coalescence of small
bubbles to the nose region of the following plug bubble as shown in Fig. 10 (b)-(d) and (e').
The mechanism of bubble detachment and attachment from the elongated plug bubble
observed in the present experiments for horizontal flow condition is quite different from
those reported for vertical flow condition. In vertical flow condition, a trailing bubble is
accelerated in the wake region of the proceeding bubble and the larger bubble overlaps the
smaller bubble in plug flow. This mechanism reported as drag-induced coalescence in
literature (Wu et al., 1998, and Kim, 1999) is not observed in the present experiments on
horizontal configuration. Thus the main sources of erosion-corrosion of protective films
inside horizontal pipe wall due to plug flow are: turbulent nature of gas and liquid flows,
fluctuating shear stresses near the pipe wall (due to large deviation in gas-liquid flow rates),
continuous change in the shape of the plug bubble rim, and the difference in pressure impact
on pipe wall.

4.2 Less aerated slug (LAS) flow

This sub-regime signifies the presence of lesser amount of aeration in liquid slugs. Also the
bubbles present inside the liquid slug are distinct. The aeration is due to the detachment of
small bubbles from the bubble tail. Due to higher turbulent mixing (which creates turbulent
breakup force sufficient to overcome the surface tension force) near the liquid slug front, the
bubbles from the region of the bubble tail starts to detach and then entrain into the liquid slug
front. Higher mixing is because of the turbulence in air (due to the higher ReSG) and high
momentum of liquid slug (due to the momentum transfer from the slug bubble to the liquid
slug). The entrained bubbles flowing inside the liquid slugs collapse repetitively due to the
pressure difference between these bubbles and the liquid plug. Such repeated implosion of
bubbles near the pipe wall causes cavitation erosion. It also leads to cyclic stress that can
result in surface fatigue of the pipe wall.
Images for less aerated slug flow are analysed (shown in Fig. 11) for 12 different
combinations of ReSL and ReSG towards illustrating the behaviour of slug front and slug tail.
As ReSG is increased from 1529 to 2677 (for fixed values of ReSL 2721, 10882, 21765), air
becomes more turbulent resulting in gradual increment of aeration inside the liquid slug front
as shown in Fig. 11 (a)-(d), (e)-(h), and (i)-(l). For higher ReSG (shown in Fig. 11 d, h and l),
the bubbles detached from bubble tail are larger in number and smaller in size. Thus, void
fraction gradually increases due to the expansion of the gas. Expansion of gas is because of
the reduction in pressure along the pipe length which leads to increase in momentum near the
liquid slug front. Higher momentum of liquid slug is responsible for higher turbulent mixing
inside the liquid slug front. The turbulent mixing increases the entrainment of gas at lower
cross-section of the pipe and creates more uniformity in the flow of bubbles near the slug
front (shown in Fig. 11 d, h and l). Due to buoyancy, bubbles near the lower pipe wall at the
liquid slug front (shown in all cases of Fig. 11 at slug front) moves towards the top of the
pipe and then they travel along the upper periphery of the pipe wall as shown near the slug
tail for all cases in Fig. 11. Growth of bubbles near the upper periphery of the pipe wall can
enhance the mechanism of sudden implosion and coalescence of bubbles and can cause
severe cavitation erosion. The dynamic forces associated with such violent collapse of
bubbles near the pipe wall, lead to high frequency fatigue-stress damage of piping material.
Due to the reduction of wake entrainment in horizontal flow (compared to vertical flow
as discussed in section 4.1), bubbles inside the liquid slug front cannot travel with the same
momentum as that of liquid slug. Such retardation in flow of bubbles inside the liquid slug
leads to the coalescence of these bubbles with following slug bubble nose. Such coalescence
of small bubbles with the following slug bubble are not observed always. Sometime these
small bubbles detached from bubble tail are swept around the slug bubble nose and reach the
bottom interface of slug bubble as shown in Fig. 11 (c), (d), (j), (h), (k) and (l). This is due to
stable interface of slug nose and the small bubbles such that the inertia of relative motion is
not sufficient to overcome the surface tension force which leads to coalescence of bubbles.
Many times, the small bubbles inside the liquid slug coalesce and forms very large bubble
known as plug bubble. However, such larger bubbles (pocket of air) are not captured here due
to the size of the viewing/capturing window. Such larger bubbles then collapse near the pipe
surface and results in pressure spikes which may erode the pipe material in their vicinity.
With increase in ReSL (10882 and 21765), bubbles inside the liquid slug front are
advected due to the higher momentum of liquid. These bubbles agglomerate near the liquid
slug front as shown in Fig. 11 (e)-(l). On the other hand, liquid level in the mixing zone rises
due to increase in ReSL and leads to the formation of slugs near the mixing zone (similar
observation as reported in plug flow). This process enhances the rapid formation of slugs.
Sudden increment in slug frequency (exact slug frequency values are reported in section 6.1
for different inflow conditions) creates enormous deviations in gas and liquid flow rates
along the pipe length. These deviations in flow rates can cause for strong Shear Stress
Induced Erosion inside the piping system. Moreover, the liquid slugs travel at much higher
velocity than encountered in normal liquid flows (exact values of liquid slug velocity are
reported in section 6.2 at different inflow conditions). Such high velocity liquid slug with
aeration can cause high turbulence and shear force at the pipe wall and thus enhance the
corrosion rate and destruction of the protective film. Jepson (1989) and Postlethwaite and
Nesic (2000) also suggested that the flow velocity of liquid slug is a drastic tool for erosion-
corrosion rates leading to the damage of the protective film inside the piping systems.

4.3 Highly aerated slug (HAS) flow

Less aerated flows tend to get highly aerated as the ReSG is increased (> 2677). Large
numbers of bubbles are observed to get entrained into the liquid slug front from the bubble
tail as shown in Fig. 12. This is due to higher air turbulence (dominant turbulent breaking
force over the surface tension). Also higher turbulent mixing near the liquid slug front
induces swirls/eddies or secondary flow inside the liquid. Based on the motion (or pathlines)
of the bubbles (near liquid slug front), eddies are identified and encircled in Fig. 12. These
eddies grow larger in size as the flow progresses (discussed later using Fig. 13). HAS flows
are mainly characterized by the presence of such eddies inside the liquid slug front. Such
secondary vortices (eddies) can significantly augment the erosion-corrosion in piping
systems. Crawford et al. (2007) has reported that the secondary flows in piping components
like elbow induce a pressure drop along the elbow wall which significantly increases the wall
mean and oscillatory shear stresses. With further increase in ReSG (beyond 3824), liquid slug
becomes more aerated and is found to be very chaotic, as shown in Fig. 12 (c)-(e), (h)-(j) and
(m)-(o). A continuous thin film of bubbles is always observed in liquid slug at the top of the
pipe (Fig. 12 (a)-(o)). Such thin film of bubbles is created by agglomeration of small bubbles
dispatched from the bubble tail. Due to the presence of large number of such small bubbles
near the upper pipe wall, the rate of implosion and coalescence of bubbles near the upper
periphery of the pipe is drastically increased. This can lead to severe cavitation erosion.
At higher ReSL (> 10882) and ReSG (> 4588), liquid slugs become smaller in size but
higher in momentum; also referred as slug precursors or pseudo slugs. Liquid level behind
the slug tail increases due to higher ReSL. The interface of liquid film is observed to be wavy
due to higher ReSG as shown in Fig. 12 (d), (e), (i), (j), (n) and (o). This residual wave is
overtaken and consumed by subsequent slugs and the number of slugs tends to increase. This
leads to significant increase in slug frequency. Such higher frequency pseudo slugs can lead
to sever shear stress induced erosion in the piping systems. In addition, these pseudo slugs
contain very large number of non-uniform bubbles (Fig. 12 (i), (j), (n) and (o)) which
augment the size of eddies and their momentum. These eddies generate secondary current in
direction opposite to the liquid slug flow resulting in high pressure spikes which can cause
vibration and jigging at locations through which such slugs pass.
Larger eddies of liquid along with large number of bubbles form the liquid droplets near
the liquid slug front. These liquid droplets strike with the pipe wall due to higher turbulence
near the liquid slug front. Strong impingement of liquid slugs or droplets to the pipe internal
surface can cause Liquid Impact Induced Erosion to piping systems. This phenomenon can
cause significant distortion to the piping system (Hinze, 1955 and Ahmed, 2012) and can lead
to higher-pressure losses in the system. Moreover, such liquid impact induced erosion along
with cavitation erosion (induced by strong bubble impingement) also referred as deformation
erosion which can erode the piping material by continued deepening of discrete sharp edged
pits (suggested by Postlethwaite and Nesic, 2000). Multiple impact of liquid droplets may
start the material removal process from internal surface of the pipe as reported by Schmitt
(1980). Furthermore, high momentum pseudo slugs can lead to Flow Accelerated Corrosion
(FAC) may cause thinning of the protective oxide layer of piping component and can create
fracture in the pipe. This may lead to catastrophic failures of system components and may
also result in serious fatalities as was reported earlier by Wood et al. (2013) and Ahmed
(2012).
Growth of eddies inside the liquid slug front captured in the present experiments are shown in
Fig. 13. Time series images of eddies inside the liquid slug front are illustrated for four
different inflow conditions (gradually increase in ReSG and keeping fixed ReSL) in Fig. 13.
Large number of bubbles, separated from the slug bubble tail mixes with eddies near the
liquid slug front and then move towards the upper periphery of the pipe (due to the
buoyancy). For all four flow conditions, at time say t = 0, small eddies are found near the
bottom of the pipe wall. As the flow progresses, they grow larger in size as shown in Fig. 13
(c), (d), (g), (h), (k), (l), (o), (p). These eddies are encircled in Fig. 13 (a)-(p) for all four flow
conditions. At high ReSG (> 3824), bubbles are not distinct inside the liquid slug due to their
random motion and large number. The mixture of eddies and bubbles leads to the formation
of liquid droplets and these droplets continuously strike the internal surface of the pipe and
then settles down. Due to this continuous (very fast) settling process of liquid droplets from
top to the bottom of the pipe, liquid (water) always remains present on the internal periphery
of the pipe. This can be described by the blurriness of the images of the liquid slug front
shown in Fig. 13 (i)-(p). Growth of eddies generate high secondary current and induces large
Turbulence Induced Collisions that can lead to very strong Liquid Impact Induced Erosion to
the internal surface of pipe.
Erosion-corrosion mechanisms reported in this section need to be analysed quantitatively in
terms of intermittent flow characteristics including plug/slug frequency, plug/slug velocity,
plug/slug length and liquid holdup.

5. Procedure adopted for measurement of intermittent flow characteristics using image

analysis

Characteristics of a liquid plug/slug (velocity, length and frequency) in every frame (image)
is determined by fixed-point data analysis approach. Procedure for fixed-point data analysis
of images is demonstrated in Fig. 14.
Liquid slug characteristics (presented in Fig. 14) are determined by considering two fixed
reference points in the field of captured images. These two points are chosen in such a
manner that they cover maximum field distance of the image. These points are marked (in
square box) on the scale (at 1.07 m and 1.42 m respectively) in Fig. 14 (a). The distance
between these two reference points is 0.35 m. Two vertical reference lines, corresponding to
these reference points are drawn (referred as reference 1 and 2) in Fig. 14 (b). Photron
FASTCAM viewer (software) facilitates the measurement of slug flow characteristics in
terms of these references suitably marking them in all the images. Liquid slug front (bubble
tail) and liquid slug tail (bubble nose) are recognized when they cross reference 1 as shown in
Fig. 14 (b) and (d) respectively. Fig. 14 (c) and (e) represent the liquid slug front and tail
crossing reference 2. Time elapsed by liquid slug front and tail for the distance travelled from
reference 1 to 2 is measured (marked in a rectangular box in Fig. 14). Axial velocity of liquid
slug front and tail are measured by the following expression:

(1)

where V refers to the axial velocity, X refers to the axial coordinates, and t refers to the time.
For the slug flow condition mentioned in Fig. 14, axial velocity of liquid slug front and tail
are calculated as 1.564 m/s and 1.573 m/s respectively. Average liquid slug velocity is
calculated as:
 (2)

The average velocity of liquid slug for the flow condition shown in Fig. 14 is 1.569 m/s. The
liquid slug length is obtained directly by measuring the position of liquid slug boundaries
(liquid slug front and tail) on the scale. For some inlet flow conditions, the liquid slug length
is larger than the viewing window of the captured image (shown in Fig. 14 (c) and (d)). In
such cases, liquid slug length is computed from the average liquid slug velocity and the time
for travel of liquid slug from reference 1 (Fig. 14 (b) and (d)) to reference 2 (Fig. 14 (c) and
(e)). It is calculated as:
(| |) (3)

For mentioned slug flow condition in Fig. 14, liquid slug length is 0.41 m. Frequency of
liquid slug is measured by counting the number of the liquid slugs (from captured images)
passing a reference point over a period of time and is expressed as:

(| |)
(4)

The method of image analysis discussed so far is highly precise and easy to process as
compared to other image processing techniques. In this regard, a comparison of present
image analysis technique with one of the most popular image processing technique reported
by Mayor et al., 2007 is presented in Table 4.

6. Quantitative analysis of intermittent flow characteristics and their influence on erosion-

corrosion in piping system

Detailed analysis of intermittent flow characteristics (including plug/slug frequency, liquid

plug/slug velocity (elongated bubble translational velocity), and liquid plug/slug length
established based on the procedure discussed in section 5 is reported in this section. These
characteristics are important mainly for two reasons. First, all existing mechanistic models
such as those of Dukler and Hubbard (1975) and Taitel and Barnea (1990) require an estimate
of these characteristics as input for pressure drop and holdup calculations. Second, and
perhaps more important for the pipeline designer, the distribution of such characteristics for
different inlet flow conditions is important for design of processing equipment and to predict
the correct erosion-corrosion rate in piping systems.

6.1 Plug/slug frequency

Plug/slug frequency is crucial since it is associated with pipeline structural instability,

flooding of downstream facilities and well head pressure fluctuation (Al-Safran, 2009). It is
also directly associated with the corrosion rate in piping system (Hill et al., 1996). It is thus
necessary to measure the plug/slug frequency for predicting the corrosion rate precisely and
to generate a correct corrosion inhibitor program (Sun and Jepson, 1992). In addition,
plug/slug frequency is needed as an input in mechanistic two-phase flow models to predict
the pressure drop and liquid holdup in pipelines (Greskovich and Shrier, 1972; Abdul-
Majeed, 2000; Al-Safran, 2009; and Al-Safran et al., 2015). In the present work, analysis is
also reported for the influence of slug/plug frequency on erosion-corrosion in piping system
for wide range of inlet flow conditions.
Fig. 15 shows the influence of ReSL and ReSG on the plug and slug frequency. Variation of
plug and slug frequency follows similar trend in plug flow, less aerated slug (LAS) flow and
highly aerated slug (HAS) regimes shown in Fig. 15 (a)-(c). However the plug and slug
frequency values vary based on the inlet flow conditions (in terms of ReSL and ReSG). At any
constant value of ReSG for all three regimes (plug, LAS, and HAS), increase in ReSL (5441 to
27206) leads to rapid formation of slug near the entry zone (mixing zone). This results in
increment in plug/slug frequency shown in Fig. 15 (a) to (c). Such increment in plug/slug
frequency causes a large deviation in gas-liquid flow rates that induces shear stress erosion in
piping structure (discussed in section 4).
On the other hand, increase in ReSG (765 to 5353) at a constant value of ReSL leads to
decrement in plug/slug frequency (see Fig. 15 (a)-(c)) due to the augmentation in gas flow
rate. Increase in gas flow rate counters the rise in liquid level, which in turn opposes the
formation of plug/slug. In general, it is found that the plug/slug frequency weakly depends on
the ReSG but is strongly influenced by ReSL (illustrated in Fig. 15 (a)-(c)).
Present experimental observations for slug frequency (in non-dimensional form) are
compared with the earlier correlations proposed by Fossa et al. (2003) and Wang et al. (2007)
in Fig. 16. Non-dimensional slug frequency variation reported by these researchers was in
terms of Strouhal number, St (fD/VSG) and liquid volume fraction, αL (VSL/VM). The present
experimental data for plug/slug frequency represented in terms of Strouhal number (St) and
liquid volume fraction (αL) shown in Fig. 16 agree closely with those reported by Fossa et al.
(2003) and Wang et al. (2007). Small deviations are due to the inclusion of plug frequency
values established in the present experiments.

6.2 Liquid plug/slug velocity (Elongated bubble translational velocity)

As explained in section 4 of the present study, slug/plug is a liquid pocket with or without
aeration. The liquid pocket flows with the gas velocity (elongated plug/slug bubble velocity).
Increase in gas velocity thus leads to augmentation in translational velocity of elongated
bubble. This translation velocity is imparted to the liquid pocket (liquid plug/slug). Thus the
liquid plug/slug travel with a nearly similar velocity as that of the elongated bubble. On the
other hand, the liquid film moves with a lower velocity corresponding to inlet liquid flow rate
(ReSL). Significant difference in flow rates of elongated bubble, liquid slug/plug and the
liquid film induces severe shear stress erosion in a piping system. In the present study, the
influence of translational velocities on erosion-corrosion phenomena is studied in detail.
Translational velocity of elongated bubbles (average of bubble tail and bubble front
velocities) are calculated by averaging the translational velocity of 50 numbers of plug/slug
for each inlet flow condition (mentioned in Fig. 9). These are represented here as function of
local mixture velocity in Fig. 17. The translational velocity of elongated bubble for different
inlet flow conditions are compared with the predicted values from the model proposed by
Bendiksen (1984) for same inlet flow conditions. It is observed that there is a clear
discrepancy between the experimental and the predicted results for lower values of mixture
velocity. However, for higher mixture velocity, good agreement between the measured values
and the predictions of Bendiksen (1984) is observed.
Fig. 18 (a)-(c) shows the bubble translational velocity (liquid plug/slug velocity) for different
inflow conditions corresponding to plug flow, LAS flow and HAS flow regime respectively.
With increase in ReSL and ReSG the velocity of liquid plug/slug increases and a similar trend
is observed for all the three regimes (see Fig. 18 (a)-(c)). For a large increase in ReSL (from
5441 to 27206), a small increment in liquid plug/slug velocity is observed for all the regimes.
On the other hand, a small increase in ReSG (at fixed ReSL), results in higher turbulent mixing
inside the liquid plug/slug (observed in section 4) which significantly augments the liquid
plug/slug velocity (see Fig. 18 (a)-(c)). This signifies that ReSG plays dominant role towards
enhancement of liquid plug/slug velocity. In addition, large aeration present inside the liquid
slug (for HAS flow regime reported in section 4) significantly augments the liquid slug
velocity compared to the other two regimes (plug and LAS). This is clearly depicted in Fig.
18 (a)-(c).As reported in section 4, two major types of erosion-corrosion (namely, flow
accelerated corrosion and liquid impact induced erosion) can result due to higher acceleration
of liquid plug/slug and very high turbulence collisions (due to higher turbulent mixing) inside
the piping systems. These two mechanisms (higher acceleration and very high turbulence
collisions) are influenced majorly by liquid plug/slug velocity.

5.3 Plug/slug liquid length

Design of plug/slug catcher or plug/slug separator is dependent on the range of the liquid
plug/slug lengths for different inlet flow conditions (Cook and Behnia, 2000). The liquid plug
and slug lengths measured in the present experiment for various inlet flow conditions are
shown in Fig. 19. It is observed that the liquid slug lengths are in the range of 12D to 24D
(shown in Fig. 19) which is closer to the observation by earlier researchers (Dukler and
Hubbard, 1975; Andreussi and Bendiksen, 1989; Nydal et al., 1992; Wang et al., 2007;
Carneiro et al., 2011). For plug flow, the liquid plug length is in the range of 8D to 24D. The
measured plug/slug lengths are the average values for 50 plugs/slugs for each flow condition.

Fig. 20 (a)-(b) shows the variation of minimum and maximum value of plug/slug length as
function of inlet flow conditions (mixture velocity). Even though the liquid plugs/slugs are
observed to flow with constant velocity for a given inlet flow condition, significant change in
their lengths has been noticed. This is due to the fact that liquid level behind the liquid
plug/slug suddenly drops once a plug/slug has passed. Before the equilibrium (original) level
of liquid film is achieved, another plug/slug is observed to be formed with smaller liquid
plug/slug length (due to the lack of the liquid volume). This results in sudden changes in
liquid slug/plug length in the intermittent flow regime. Continuous fluctuation in liquid
plug/slug length causes difficulty in prediction of correct erosion-corrosion inhibitor program
and designing precise plug/slug separator in piping systems. Minimum value of slug/plug
length is observed in the range of 4D to 14D and maximum value is found in the range of
14D to 38D for intermittent flow regime as shown in Fig. 20. Maximum and minimum values
are important to be known since they decide the design of slug catchers/separators.
Influence of ReSL and ReSG on liquid length for the plug, LAS, and HAS flow regimes is
shown in Fig. 21 (a)-(c) respectively. With increase in ReSL (for the same ReSG), the ratio of
liquid to gas volume in the plug/slug unit increases. This is depicted by the increase in film
height and reduction in mass of gas contained by elongated bubble. Such increment in liquid
film height causes rapid formation of plug/slug for a given flow condition. This in turn
increases the frequency of slug/plug (discussed in section 6.1) and reduces the liquid
plug/slug length (as shown in Fig. 21). On the other hand, at constant ReSL, increase in ReSG
(or transition from plug to LAS and LAS to HAS) results in increase in liquid slug length due
to higher bubble translational velocity which augments the picking rate of liquid from the
preceding slug. The elongated bubble length also increases with increase in ReSG which
causes reduction in slug/plug frequency (discussed in section 6.1). Such degradation and
elongation of liquid plug/slug lengths for different inlet flow conditions directly influences
the erosion-corrosion rate in piping systems (Jepson, 1989).

7. Development of erosion-corrosion map for intermittent flow regime

The control of erosion-corrosion in petroleum industry can be expensive and time consuming
due to the unplanned nature of stoppages and high maintenance costs (Wood et al., 2013;
Okonkwo and Mohamed, 2014). Proper design and safer operation of pipelines to avoid such
erosion-corrosion are of great interest to the broader petroleum community. The erosion-
corrosion phenomena are influenced mainly by fluid properties, flow conditions and flow
characteristics. Based on the flow visualisation (reported in section 4) and quantitative
analysis of intermittent flow characteristics presented so far, a non-dimensional erosion-
corrosion map in terms of varying inlet flow conditions is developed in this paper. This map
will provide guidelines for safer monitoring of inflow conditions during pipeline operations
to avoid erosion-corrosion problems (identified in section 4). The non-dimensional erosion-
corrosion map so developed (shown in Fig. 22) demarcates the transition boundaries between
four major types of erosion-corrosion phenomena: Shear stress induced erosion (SSIE),
cavitation erosion (CE), liquid impact induced erosion (LIIE), and flow accelerated Corrosion
(FAC).

Shear stress induced erosion (SSIE) is formed due to large and frequent deviation in gas-
liquid flow rates (reported in section 4). Higher frequency of smaller plug/slug (flowing with
high velocity) influences such deviation in flow rates (discussed in section 6.1). Based on
plug/slug frequency (> 1.35 Hz) and velocity (> 1 m/s) for different inlet flow conditions,
transition boundary AA' for SSIE is constructed on the regime map shown in Fig. 22.
Cavitation erosion (CE) is associated with LAS and HAS flow regime due to the collapse of
bubbles inside the liquid slug near the pipe wall. Aeration inside the liquid slug increases
greatly above 1.5 m/s of mixture velocity of the phases (or beyond 1.8 m/s of liquid slug
velocity). Such increment in aeration inside the liquid slug significantly augments the
mechanism of bubble implosion, coalescence, breakup, and agglomeration due to higher
turbulent mixing inside the liquid slug. This in turn can cause severe cavitation erosion (CE)
in piping material (reported in section 4). Based on the bubble translational velocity (> 1.8
m/s) or inlet mixture velocity (> 1.5 m/s), transition line BB' for CE is constructed in Fig. 22.
High velocity of liquid slug (> 2.2 m/s) is responsible for flow accelerated corrosion (FAC).
Based on the liquid slug velocity, transition line CC' for FAC is constructed in Fig. 22.
Moreover, very high liquid slug velocity (> 3.5 m/s) can induce liquid impact erosion due to
strong turbulent induced collisions inside the liquid slug (reported in section 4). Transition
line DD' for liquid impact induced erosion (LIIE) is constructed using bubble translational
velocity (> 3.8 m/s) in Fig. 22.
Four distinct transition boundaries (AA', BB', CC' and DD') divide the intermittent flow
regime in eight different regions as shown in Fig. 23. These regions provide information
about the perilous range of inlet flow conditions (ReSL and ReSG) for which typical kind of
erosion-corrosion phenomena may occur. This erosion-corrosion map (presented in Fig. 23)
can be a great asset for pipeline designer to minimize the economic loss due to erosion-
corrosion.
Region 1 (area covered in AEB of Fig. 23) is a safer zone in context to probability of erosion-
corrosion occurring in the pipe. This means that if the operation of the pipe is for the inflow
conditions corresponding to the values of ReSL and ReSG within this region (area inside AEB
of Fig. 23) it is highly possible that there will be no erosion-corrosion in the pipe. Region 2
covered within area AEB' in Fig. 23 represents the regime which is prone to shear stress
induced erosion (SSIE). This is a type of erosion, which is due to enormous deviation of gas,
and liquid flow rates associated with plug flow. As discussed earlier in this section, cavitation
erosion (CE) can form only in LAS and HAS flow regime. Accordingly, region 3 within area
BCEF shown in Fig. 23 represents the regime for CE. Region 4 within area of B'EFC' (within
LAS zone) in Fig. 23 is a regime, operation of pipe inflow conditions in which can result in
two types of erosions: CE and SSIE. Region 5 (area covered in C'FGD' of Fig. 23) is a strong
zone for three types of erosion-corrosion phenomena: CE, SSIE and FAC. These are
associated with LAS and HAS flows. Region 6 (area covered in FCDG of Fig. 23) is a strong
zone for CE and FAC, caused majorly due to HAS flow. Region 7 (area covered in DGA' of
Fig. 23) is a strong zone for CE, FAC and LIIE, associated with HAS flow. Region 8 (area
covered within D'GA' of Fig. 23) is a zone of inflow conditions which can lead to all kinds of
erosion-corrosion phenomena and is thus most perilous in context to pipeline operation. The
erosion-corrosion map (Fig. 23) devised in the present research for different inlet flow
conditions can thus immensely assist petroleum and nuclear engineers for safer operation of
pipelines avoiding or minimizing erosion and corrosion of pipes.

8. Summary and Conclusions

Despite the prevalence of catastrophic failures in petroleum and nuclear industries, limited
studies are available in literature dealing with erosion-corrosion due to interfacial structure of
intermittent flow. In this direction, flow visualization study is carried out to examine the
internal structure of intermittent flow sub-patterns and to determine the associated erosion-
corrosion phenomena inside the piping systems. Three major sub-patterns of intermittent flow
(Plug flow with or without bubble rim, LAS flow and HAS flow) are investigated. Based on
their interfacial dynamics, four major erosion-corrosion mechanisms are predicted inside the
piping systems: Shear Stress Induced Erosion (SSIE), Cavitation Erosion (CE), Liquid
Impact Induced Erosion (LIIE), and Flow Accelerated Corrosion (FAC). Major conclusions
for all three sub-regimes are presented in following manner:
 Expansion, contraction and breakup of the elongated plug bubble rim leads to
initiation of corrosion in pipe. Reduction in length of plug and increment in number of
plugs at higher ReSL (> 10882) creates enormous deviations in gas-liquid flow rates
and can cause shear stress induced erosion inside the piping material. On other hand,
increase in ReSG (> 1200) leads to the entrainment of aeration inside the liquid slug
front and results into transition of plug flow to less aerated slug flow.
 Higher turbulent mixing near the liquid slug front (due to increase in ReSG > 1200)
augments the entrainment of gas bubbles into the slug front which travel to the upper
pipe wall due to buoyancy. Growth and collapse of such bubbles near the pipe wall
results in pressure spikes which may erode the pipe material in its vicinity (called
cavitation erosion). Due to the absence of wake entrainment, such bubbles coalesce to
form larger bubbles (called plug bubble) which can collapse near the pipe wall as the
flow progresses and can enhance cavitation erosion to the piping material.
 Swirls/eddies in the highly aerated liquid slug front (ReSG > 2677) oppose the motion
of the liquid slug and create very high pressure spikes causing vibration and jigging
on the piping system. Highly aerated liquid slugs (at ReSG > 4588) are very chaotic
and smaller in size (referred as pseudo slugs or slug precursor). Small liquid droplets
from these pseudo slugs strike the pipe wall strongly and cause liquid impact induced
erosion inside the piping system. Multiple impact of such liquid droplets can lead to
onset of material removal process from internal surface of the pipe.
 Very high momentum of pseudo slugs can cause fracture to the internal structure of
pipe or may cause thinning of the protective oxide layer of piping components or
system. This phenomena is called as flow accelerated corrosion (FAC). Very high
ReSL (> 10882) and ReSG (> 4588) leads to rapid formation of pseudo slugs. This
creates large deviations in gas and liquid flow rates, which can cause severe shear
stress induced erosion inside the piping systems.
 Growth of eddies inside the highly aerated liquid slugs generates secondary currents
and induces large turbulence induced collisions which can lead to very strong liquid
impact induced erosion to the internal surface of pipe.
 The minimum liquid plug/slug lengths are in the range of 4D to 14D while the
maximum values are in the range of 14D to 38D respectively for the entire
intermittent regime of flow. Estimation of minimum and maximum slug/plug length
helps in design of slug/plug catchers.
Based on the database of intermittent flow characteristics and visual observations of
intermittent flow sub-patterns, an erosion-corrosion map is devised for predicting four major
types of erosion-corrosion in pipes in terms of ReSG and ReSL (inlet flow conditions). Eight
distinct regimes of erosion-corrosion are identified in this map. This erosion-corrosion map
can be used by petroleum and nuclear engineers to prepare more reliable scope for plant
inspection in order to avoid intermittent flow associated erosion-corrosion in piping systems.
Acknowledgement
The authors gratefully acknowledge the financial assistance provided by Science and
Engineering Research Board (SERB), Department of Science and Technology (DST), India
(sanction letter no SB/S3/MIMER/0111/2013 dated 23-05-2014) and Center of Excellence
(COE) on Water Resources and Flood Management (sanction letter no. NPIU/TEQIP-
II/FIN/47 dated 21-06-2013), Sardar Vallabhbhai National Institute of Technology, India.

References

Abdul-Majeed, G., 2000. Liquid slug holdup in horizontal and slightly inclined two- phase
slug flow. J. Pet. Sci. Eng. 27, 27–32.
Ahmed, W.H., 2012, Nuclear Power- Practical Aspects. InTech, New York, USA, Chapter 6,
153–178.
Ahmed, W.H., Bello, M.M., Al-Sarkhi, A., El Nakla, M., Badr, H.M., 2014. Experimental
investigation of flow accelerated corrosion under two-phase flow conditions. Nuclear
Engineering and Design 267, 34–43.
Andritsos, N., Williams, L., Hanratty, T.J., 1989. Effect of liquid viscosity on the stratified-
slug transition in horizontal pipe flow. Int. J. Multiphase Flow 15, 877–892.
Al-Safran, E., 2009. Prediction of slug liquid holdup in two-phase horizontal pipes. ASME J.
Energy Res. Technol. 131, 1–8.
Al-Safran, E., Kora, C., Sarica, C., 2015. Prediction of slug liquid holdup in high viscosity
liquid and gas two-phase flow in horizontal pipes. J. Pet. Sci. Eng. 69, 566–575.
Alssayh, M., Addali, A., Mba, D., Naid, A., 2013. Detecting and measuring slug velocity in
two-phase flows horizontal pipes. 1st Euro-Mediterranean Conference on Structural
Dynamics and Vibroacoustics, Marrakech, Morocco.
Alves, I.N., Shoham, O., 1993. Drift velocity of elongated bubbles in inclined pipes.
Chemical Engineering Science 48 (17), 3063-3070.
Andreussi, P., Bendiksen, K., 1989. An investigation of void fraction in liquid slugs for
horizontal and inclined gas-liquid pipe flow. Int. J. Multiphase Flow 15 (6), 937–946.
Bendiksen, K.H., 1984. An experimental investigation of the motion of the long bubbles in
inclined tubes. Int. J. Multiphase Flow 10 (4), 467–483.
Bozzini, B., Ricotti, M.E., Boniardi, M., Mele C., 2003. Evaluation of erosion-corrosion in
multiphase flow via CFD and experimental analysis. Wear, 14th International Conference
on Wear of Materials 255 (1-6), 237–245.
Carneiro, J.N.E., Fonseca Jr, R., Ortega, A.J., Chucuya, R.C., Nieckele, A.O., Azevedo,
L.F.A., 2011. Statistical characterization of two-phase slug flow in a horizontal pipe.
Journal of the Brazil Society of Mech. Sci. and Engg. 33, 251–258.
Chang, Y.S., Song, K.H., Lee, S.M., Choi, J.B., Kim, Y.J., 2006. Key Eng. Mater. 321–323,
670–673.
Chexal, B., Horowitz, J., Jones, R., Dooley, B., Wood, C., Bouchacourt, M., Remy, F.,
Nordmann, F., Paul, St.P., 1996. Flow accelerated corrosion in power plants, Electric
Power Research Institute Report No. TR-106611.
Cook, M., Behnia, M., 1997. Film profiles behind liquid slugs in gas-liquid pipe flow.
AICHE J. 43 (9), 2180–2186.
Cook, M., Behnia, M., 2000. Slug length prediction in near horizontal gas–liquid intermittent
flow. Chem. Eng. Sci. 55, 2009–2018.
Crawford, N.M., Cunningham, G., Spence, S.W.T., 2007. An experimental investigation into
the pressure drop for turbulent flow in 90° elbow bends. Proc. of the Institution of
Mechanical Engineers, Part E, Journal of Process Mech. Engg. 221 (2), 77–88.
Czapp, M., Muller, C., Fernandez, P.A., Sattelmayer, T., 2012. High-speed Stereo and 2D
PIV measurements of two-phase slug flow in a horizontal pipe. 16th International
Symposium on Applications of Laser Techniques to Fluid Mechanics, Lisbon, Portugal.
De Henau, V., Raithby, G.D., 1995. A transient two-fluid model for the simulation of slug
flow in pipelines: I. Theory. Int. J. Multiphase Flow 21(3), 335–349.
Dukler, A.E., Hubbard, M.G., 1975. A model for gas–liquid slug flow in horizontal and near
horizontal tubes. Ind. Eng. Chem. Fundam. 14, 337–345.
Fabre, J., and Liné, A., 1992. Modelling of two phase slug flow. Ann. Rev. Fluid Mech. 24,
21–46.
Ferng, Y.M., 2008. Prediction local distribution of erosion-corrosion wear sites for the piping
in the nuclear power plant using CFD models. Ann. Nucl. Energ. 35 (2), 304–313.
Ferng, Y.M., Ma, Y.P., Chung, N.M., 2000. Application of local flow models in predicting
distributions of erosion–corrosion locations. Corrosion 56 (2), 116.
Fossa, M., Guglielmini, G., Marchitto, A., 2003. Intermittent flow parameters from void
fraction analysis. Flow Meas. Instrum. 14, 161–168.
Ghajar, A. J., Tang, C. C., 2007. Heat transfer measurement, flow pattern maps and flow
visualization for non-boiling two-phase flow in horizontal and slightly inclined pipe.
Heat Transfer Engineering 28 (6), 525–540.
Gregory, G.A., Scott, D.S., 1969. Correlation of liquid slug velocity and frequency in
horizontal cocurrent gas–liquid slug flow, AIChE J. 15, 833–835.
Greskovich E.J., Shrier A.L., 1972. Slug frequency in horizontal gas-liquid slug flow. Ind.
Eng. Chem. Process. Des. Dev. 11(2), 317–318.
Hanyang, G., Liejin, G., 2008. Experimental investigation of slug development on horizontal
two-phase flow. Chinese Journal of Chemical Engineering 16 (2), 171–177.
Heywood, N., Richardson, J., 1979. Slug flow of air–water mixtures in a horizontal pipe:
determination of liquid holdup by gamma-ray absorption. J. Chem. Eng. Sci. 34, 17–30.
Hill, T., Fairhurst, C., Nelson, C., Becerra, H., Bailey, R., 1996. Multiphase production
through hilly-terrain pipelines in Cusiana oilfield, Colombia. SPE 36606, Presented at
SPE ATCE, Denver, CO.
Hinze, J.O., 1955. Fundamentals of the hydrodynamic mechanism of splitting in dispersion
processes. AIChEJ. 1 (3), 289–295.
Hubbard, M. G., 1965. An analysis of horizontal gas-liquid slug flow. Ph.D. Thesis, Univ. of
Houston.
Jepson, W.P., 1989. Modelling the transition to slug flow in horizontal conduit. Canadian
Journal of Chemical Engineering 67, 731–740.
Kadri, U., Zoeteweij, M.L., Mudde, R.F., Oliemans, R.V.A., 2009. A growth model for
dynamic slugs in gas–liquid horizontal pipes. Int. J. Multiphase Flow 35, 439–449.
Kaul, A., 1995. Study of slug flow characteristics and performance of corrosion inhibitors, in
multiphase flow, in horizontal oil and gas pipelines, M.S. Thesis, Ohio University,
Athens, Ohio.
Keating, A., Nesic, S., 1999. Prediction of two-phase erosion-corrosion in bends. In: Second
International Conference on CFD in the Mineral and Process Industries, CSIRO,
Melbourne, Australia, December 6–8.
Kim, S., 1999. Interfacial area transport equation and measurement of local interfacial
characteristics, Ph.D. Thesis. Purdue University.
Kim, S., Park, J.H., Kojasoy, G., Kelly, J.M., Marshall, S.O., 2007. Geometric effects of 90-
degree elbow in the development of interfacial structures in horizontal bubbly flow.
Nuclear Engineering and Design 237, 2105–2113.
Kjeldby, T.K., Henkes, R.A.W.M., Nydal, O.J., 2013. Lagrangian slug flow modelling and
sensitivity on hydrodynamic slug initiation methods in a severe slugging case. Int. J.
Multiphase Flow 53, 29–39.
Kuo-Tong, M.A., Ferng, Y.M., May, P., 1998. Numerically investigating the influence of
local flow behaviors on flow-accelerated corrosion using two-fluid equations. Nuclear
Technology 123, 90–102.
Kvernvold, O., Vindoy, V., Sontvedt, T., Saasen, A., Selmer-olsen, S., 1984. Velocity
distribution in horizontal slug flow. Int. J. Multiphase Flow 10 (4), 441–457.
Lewis, S., Fu, W.L., Kojasoy, G., 2002. Internal flow structure description of slug flow-
pattern in a horizontal pipe. Int. J. Heat and Mass Transfer 45, 3897–3910.
Li, G., Wang, Y., He, R., Cao, X., Lin, C., Meng, T., 2009. Numerical simulation of
predicting and reducing solid particle erosion of solid–liquid two-phase flow in a choke.
Petroleum Science 6 (1), 95–97.
Lin, P.Y., 1985. Flow regime transitions in horizontal gas-liquid flow. Ph.D. Dissertation,
University of Illinois, Urbana.
Mayor, T.S., Pinto, A.M.F.R., Campos, J.B.L.M., 2007. An image analysis technique for the
study of gas-liquid slug flow along vertical pipes–associated uncertainty. Flow
Measurement & Instrumentation 18, 139–147.
Nydal, O.J., Pintus, S., Andreussi, P., 1992. Statistical characterization of slug flow in
horizontal pipes. Int. J. Multiphase Flow 18, 439–453.
Okonkwo, P., Mohamed, A.M.A., 2014. Erosion-corrosion in oil and gas industry: A review.
Int. J. Metallurgical and Materials, 4 (3) 7–28.
Postlethwaite, J., Nesic, S., 2000. Erosion-corrosion in single and multi-phase flow - Chapter
15, Uhlig’s Corrosion Handbook, Second Edition, John Wiley & Sons, Inc.
Poulson, B., 1990. Mass transfer from rough surfaces. Corrosion Science 30 (6/7), 743–746.
Poulson, B., 1999. Complexities in predicting erosion corrosion. Wear 233–235, 497–504.
Prabha, S.S., Rathish, R.J., Dorothy, R., Brindha, G., Pandiarajan, M., Al-Hashem, A.,
Rajendran, S., 2014. Corrosion problems in petroleum industry and their solution. Eur.
Chem. Bull. 3 (3), 300–307.
Reis, E., Goldstein Jr. L., 2005. A non-intrusive probe for bubble profile and velocity
measurement in horizontal slug flows, Flow Measurement and Instrumentation 16, 229–
239.
Schmitt, G. F., 1980. Liquid and solid particle impact erosion. Wear Control Handbook,
ASME, 231–282.
Sharma, S., Lewis, S., Kojasoy, G., 1998. Local studies in horizontal gas-liquid slug flow.
Nuclear Engineering and Design 184, 305–318.
Shawki, S., 2011. Internal corrosion of petroleum pipelines. J. Engg. Tech. 1 (3), 112–117.
Shirazi, S.A., McLaury, B.S., Shadley, J.R., Rybicki, E.F., 1995. Generalization of the API
RP14E guideline for erosive services. Journal of Petroleum Technology 47 (8), 693–698.
Sun, J., Jepson, P., 1992. Slug flow characteristics and their effect on corrosion rates in
horizontal oil and gas pipelines. SPE 24787, Presented at SPE ATCE, Washington, DC.
Taitel, Y., Barnea, D.A., 1990. A consistent approach for calculating pressure drop in
inclined slug flow. Chemical Engineering Science 45, 1199–1206.
Taitel, Y., Dukler, A.E., 1976. A model for predicting flow regime transitions in horizontal
and near horizontal gas–liquid flow. AIChE J. 22 (1), 47–55.
Thaker, J., Banerjee, J., 2015. Characterization of two-phase slug flow sub-regimes using
flow visualization. J. Pet. Sci. Eng. 135, 561–576.
Ujang, P.M., Hale, C.P., Lawrence, C.J., Hewitt, G.F., 2006. Slug initiation and evolution in
two-phase horizontal flow. Int. J. Multiphase Flow 32, 527–552.
United Kingdom Accreditation Service (UKAS), 2007. The expression of uncertainty and
confidence in measurement, M3003, Edition 2.
Van Hout, R., Barnea, D., Shemer, L., 2001. Evolution of statistical parameters of gas–liquid
slug flow along vertical pipes. Int. J. Multiphase Flow 27 (9), 1579–1602.
Vaze, M.J., Banerjee, J., 2011. Experimental visualisation of two-phase flow patterns and
transition from stratified to slug flow. Proc. IMechE, Part C J. Mech. Eng. Sci. 225, 382–
389.
Vaze, M.J., Banerjee, J., 2012. Prediction of liquid height for onset of slug flow, Can. J.
Chem. Engg. 90(5), 1295–1303.
Vermeulen, L.R., Ryan, J.T., 1971. Two-phase slug flow in horizontal and inclined tubes.
Canadian Journal of Chemical Engineering 49, 195–201.
Wang, X., Guo, L., Zhang, X., 2007. An experimental study of the statistical parameters of
gas-liquid two-phase slug flow in horizontal pipeline. Int. J. Heat Mass Transfer 50,
2439–2443.
Woods, B.D., Hanratty, T.J., 1999. Influence of Froude number on physical processes
determining frequency of slugging in horizontal gas-liquid flows. Int. J. Multiphase Flow
25 (6-7), 1195–1223.
Wood, M.H, Vetere Arellano, A.L., Van Wijk, L., 2013. Corrosion related accidents in
petroleum refineries: lessons learned from accidents in EU and OECD Countries. JRC
Scientific and Policy Reports, EUR 26331 EN.
Wu, Q., Kim, S., Ishii, M., Beus, S.G., 1998. One-group interfacial area transport in vertical
bubbly flow. Int. J. Heat Mass Transfer 41, 1103–1112.
Yu, S., Li, S., Zeng, H., Li, J., Liang, R., 2009. Predicting corrosion remaining life of
underground pipelines with a mechanically based probabilistic model. J. Pet. Sci. Eng.
65, 162–166.
Yuh, M.F., Yung, S.T., Bau, S.P., Long, W.S., 2008. A two-phase methodology to predict
FAC wear sites in the piping system of a BWR. Nuclear Engineering and Design 238,
2189–2196.

Video.1 Video showing breakup of the plug bubble rim

Video.2 Video showing plug bubble rim contraction

Video.3 Video demonstrating plug bubble rim expansion

Fig. 1 Photograph of the experimental test facility for two-phase flow (TPFTR)
Fig. 2 Schematic diagram of two-phase flow test rig (TPFTR)
Fig. 3 Flowchart of supervisory control and data acquisition system
Fig. 4 Two-phase flow mixture (a) Photograph (b) Cross-sectional view (c) 3D view
Fig. 5 High speed photography system
Fig. 6 Internal structures of intermittent flow sub-patterns
Fig. 7 Sub-regime map for intermittent flow in horizontal pipe
Fig. 8 Validation of present experimental results with the earlier experimental results of Lin
(1985) and Andritsos et al. (1989) for onset of slugging

Fig. 9 Data points considered for analysis of interfacial structure of intermittent flow sub-
patterns
Fig. 10 Plug flow for different inlet conditions
Fig. 11 Less aerated slug flow for different inlet flow conditions
Fig. 12 Highly aerated slug flow for different inlet flow conditions
Fig. 13 Growth of eddies near the liquid slug front with increase in ReSG.
Fig. 14 Representation of the camera field view in fixed-point analysis during the passage of
a liquid slug
Fig. 15 Influence of ReSL and ReSG on plug/slug frequency: (a) Plug flow; (b) LAS flow; and
(c) HAS flow
Fig. 16 Comparison of present experimental slug frequency data with the correlations of
Fossa et al. (2003) and Wang et al. (2007)
Fig. 17 Translational velocity of elongated bubble as function of mixture velocity:
Comparison of present measurements to the data of Bendiksen (1984)
Fig. 18 Influence of ReSL and ReSG on bubble translational velocity: (a) Plug flow; (b) LAS
flow; and (c) HAS flow
 Fig. 19 Variation of liquid plug/slug length with mixture velocity

Fig. 20 Variation of minimum and maximum liquid plug/slug lengths with mixture velocity
Fig. 21 Influence of ReSL and ReSG on liquid length: (a) Plug flow; (b) LAS flow; and (c)
HAS flow
Fig. 22 Transition boundaries for four major types of erosion-corrosion phenomena
associated with intermittent flow regime
 Fig. 23 Erosion-corrosion map for intermittent flow regime

Table 1 Summary of experimental studies on slug flow in horizontal macro-scale pipe

Authors/ Test rig
Fluid Major Research Contribution
Year diameter
Hubbard Air- 35.1 First comprehensive experimental investigations on slug
(1965) Water mm frequencies was performed.
Gregory Liquid slug velocities and frequencies have been
CO2- 19.05
and Scott measured and correlation for slug frequency was
Water mm
(1969) developed.
Vermeulen Air- 12.7 Slug characteristics (including, pressure gradient,
and Ryan Water mm pressure fluctuations, slug velocity and slug frequency)
 (1971) for horizontal and inclined flow were measured.
Heywood
Pressure gradient and time-averaged values of liquid
and Air-
42 mm holdup for slug flow have been measured using gamma-
Richardson Water
ray absorption technique.
(1979)
Velocity distribution in slug flow was measured by
Kvernvold
newly established experimental technique. This
et al. NO2-Oil 24 mm
technique consisted LDV, an optical two phase probe
(1984)
with a signal conditioner and an electronic gate facilities.
53 mm
Nydal et Air- Mean slug characteristics (including slug liquid length,
and 93
al. (1992) Water holdup and velocity) have been determined.
mm
Alves et al. Air- An inviscid flow theory was applied to obtain the drift
51 mm
(1993) Kerosene velocity of elongated slug bubbles.
Cook and 32 mm A comprehensive one-dimensional fully coupled
Air-
Behnia and 50 hydrodynamic treatment of the film profiles behind
Water
(1997) mm liquid slugs was reported.
Woods and
Air- 76.3 Effect of liquid Froude number on the physical processes
Hanratty
Water mm for forming slugs was demonstrated.
(1999)
Cook and Rate of collapse of short slugs as a function of slug
Air-
Behnia 50 mm length was determined by measuring the statistical
Water
(2000) distribution of slug lengths.
van Hout 24 mm The translational velocities of elongated bubbles in
Air-
et al. to 54 continuous slug flow were measured for various flow
Water
(2001) mm rates, pipe inclinations and pipe diameters.
Internal flow structure of a horizontal slug flow-pattern
Lewis et Air- 50.3
(especially velocity profile in liquid slug) was discussed
al. (2002) Water mm
using hot-film anemometry technique.
Authors/ Test rig
Fluid Major Research Contribution
Year diameter
A new procedure was proposed to characterize the
40 mm
Fossa et al. Air- intermittent flow through the statistical analysis of the
and
(2003) Water instantaneous cross-sectional averaged void fraction by
60 mm
means of ring impedance probes.
Reis and Mean velocity of elongated bubbles in slug flow was
Air-
Goldstein 34 mm measured using newly established capacitance probe
Water
(2005) technique.
Ujang et Air- Initiation and the subsequent evolution of hydrodynamic
78 mm
al. (2006) Water slugs were studied in detail.
Statistical parameters (liquid slug velocities, length of
Wang et Air-
50 mm liquid slugs and elongated bubbles, and slug frequencies)
al. (2007) Water
of slug flow were measured using two-pairs of
 conductivity probs.
Hanyang
Air- Slug initiation and subsequent evolution were examined
and Liejin 50 mm
Water in detail.
(2008)
Phenomenological investigation was carried out to study
Al-Safran 50.8 the parameters controlling slug frequency in pipe flow
Air-Oil
(2009) mm and to develop a predictive empirical correlation for slug
frequency.
Carneiro et Air- Slug flow parameters were experimentally and
24 mm
al. (2011) Water numerically investigated.
Vaze and
Air- 25.4 Critical liquid height for onset of slugging was measured
Banerjee
Water mm for different inlet flow conditions.
(2012)
High speed Stereo PIV combined with LIF was
successfully applied on round pipe geometry to
Czapp et Air-
54 mm determine the instantaneous three-component water
al. (2012) Water
velocity field of slug formation process in pipe cross-
section.
This experimental study investigated the application of
Alssayh et Air-
50 mm Acoustic Emission (AE) technology for detecting slug
al. (2013) Water
velocity in two phase (gas/liquid) flow.
Thaker and Five distinct sub-regimes of slug flow were identified
Air-
Banerjee 25 mm based on the morphology of interface and elongated
Water
(2015) bubble shape geometry.

Table 2 Summary of data points considered for analysis of interfacial structure of intermittent
flow sub-regimes

Sub-regime Data points

Plug flow 14

Less aerated slug flow 20

Highly aerated slug flow 20

 Table 3 Uncertainty analysis

Standard Expanded
Parameters Parameters Uncertainty (%)
Uncertainty (%)
Pipe diameter (25 mm) 0.60 Superficial Reynolds
0.71
Mass flow rate of air 0.38 number of gas (ReSG)
Superficial Reynolds
Mass flow rate of water 0.13 0.61
number of liquid (ReSL)

Table 4 Comparison of present image analysis technique with image processing technique
reported by Mayor et al. (2007)

Image processing technique reported by

Present image analysis technique
Mayor et al., 2007

Experimental arrangement is required

(adjusting the scale with perfect
No need for any experimental arrangement
alignment over the pipe surface (without
any disturbance of pipe orientation))

Commercial tool MATLAB is essential for No need for any commercial tool for
such image processing technique analysing the captured images
Image treatment processes are required
before the image analysis (like image
loading, image conversion, image contrast No need for any image treatment
enhancement, image filtering, image processes before the image analysis
conversion to binary mode, image objects
labelling, object analysis, and image erosion)

Calibration of such process is required No need for any calibration

No need for such conversion. Present

Conversion of pixel units to real length units
technique measures the real length unit
is required
only

Very easy to measure each and every

Movement of small bubbles inside the liquid
distinct bubble movement inside the
slug are difficult to measure
liquid slug

Many times such technique is failed to

Easy to distinguish the reference of
measure the parameter when the large
 number of the bubbles are agglomerated near measurement at liquid slug front and tail
the liquid slug front and tail

Error (uncertainty) analysis of each

No need for any error (uncertainty)
measured parameters is highly essential for
analysis
accurate measurement

Highlights
 Dynamics of intermittent flow is analysed in terms of expansion, contraction,
breaking, coalescence, collapse, and collision of bubbles.
 Intermittent flow dynamics is correlated to shear stress induced erosion, cavitation
erosion, liquid impact induced erosion and flow accelerated corrosion.
 Erosion-corrosion regime map is established as a function of inlet flow conditions of
both the phases.

 Eight distinct regimes of erosion-corrosion are illustrated in erosion-corrosion map in

terms of their perilous influence.