See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/315060802

Hydrate Formation and Plugging Mechanisms in Different Gas-Liquid Flow

Patterns

Article  in  Industrial & Engineering Chemistry Research · March 2017

DOI: 10.1021/acs.iecr.6b02717

CITATIONS READS

6 78

7 authors, including:

Bohui Shi Xiaofang Lv

China University of Petroleum Changzhou University
39 PUBLICATIONS   161 CITATIONS    22 PUBLICATIONS   95 CITATIONS   

SEE PROFILE SEE PROFILE

Yang Liu Haihao Wu

China University of Petroleum China University of Petroleum, China, Beijing
17 PUBLICATIONS   36 CITATIONS    11 PUBLICATIONS   34 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The fluidization exploitation of non-diagenetic hydrate deposits View project

All content following this page was uploaded by Xiaofang Lv on 31 October 2018.

The user has requested enhancement of the downloaded file.

 Article

pubs.acs.org/IECR

Hydrate Formation and Plugging Mechanisms in Different Gas−

Liquid Flow Patterns
Lin Ding,† Bohui Shi,*,† Xiaofang Lv,†,‡ Yang Liu,† Haihao Wu,† Wei Wang,† and Jing Gong*,†
†
National Engineering Laboratory for Pipeline Safety/MOE Key Laboratory of Petroleum Engineering/Beijing Key Laboratory of
Urban Oil and Gas Distribution Technology, China University of Petroleum-Beijing, Beijing 102249, People’s Republic of China
‡
Jiangsu Key Laboratory of Oil and Gas Storage and Transportation Technology, Changzhou University, Changzhou, Jiangsu 213016,
People’s Republic of China

ABSTRACT: As oil/gas exploitation moves into deep water,

hydrate formation and plugging in flowline have been a main
concern of the flow assurance engineers. A series of experiments
were conducted in a gas-emulsion multiphase flow system using
a high pressure flow loop. The properties of hydrate
agglomeration and deposition in different flow patterns were
investigated. First, based on the hydrate chord length
distribution and the changes of slurry density, several methods
were proposed to quantitatively estimate the hydrate agglom-
eration degree and deposition degree. Second, typical results in
each flow pattern were analyzed, and the plug formation
mechanisms in each flow pattern were proposed. Then, after
comparing the results in each flow pattern, it was found that the
order of hydrate agglomeration degree from high to low is slug flow, stratified flow, bubble flow, and annular flow; and the order
of hydrate deposition degree from high to low is annular flow, slug flow, bubble flow, and stratified flow.

1. INTRODUCTION liquid multiphase flow, especially the plug mechanism in

Natural gas hydrates are complex crystalline, icelike solids different flow patterns, is an important problem to be solved.
In the past few years, some studies have been done to
formed by natural gas molecules and water under low
investigate the influence of hydrate formation on multiphase
temperature and high pressure.1 In recent years, as the
flow property. In the experimental aspect, Joshi et al.21
development tendency of petroleum industry moves to deep
observed in experiments for the first time that the formation
water, hydrates have been a major hazard to the flow assurance
of a small quantity of hydrates may immediately lead to slug
of subsea transportation systems.2 Hydrate formation in the flow onset in a system near the stratified/slug flow or bubble/
pipeline can increase the pressure loss, make the flow more slug flow transition point. This is the first evidence that hydrate
complex and unpredictable, and may even block the pipeline.3 can affect the flow pattern transition in multiphase flow
Traditional ways of hydrate prevention, including pipeline systems. Then, to further identify that whether hydrate
insulation, pressure reduction, and thermodynamic inhibitor formation influences the flow pattern transition, Lv et al.22
injection, have been used for decades in field production; but studied the flow pattern of the gas-slurry flow with hydrate
the traditional ways are of great economic cost that may particles on a high pressure flow loop with 1 in. internal
become extremely unacceptable in future field production with diameter. Their results revealed that the Mandhane flow
high water cut. Antiagglomerant is an alternative approach in pattern map could not predict the gas-slurry flow pattern
the hydrate management strategy, which allows hydrates to efficiently, and this indicated that hydrate formation did affect
form but prevents their agglomeration.4 In this condition, the the flow pattern transition apparently. However, because the
system forms a slurry flow in flowline. In this approach, the Mandhane flow pattern map has a great dependence on the
hydrate slurry must have good fluidity and must not block the experimental apparatus and materials, it is likely that the results
flowline. Therefore, to promote a better application of the of Lv et al. may not reflect the influence of hydrate formation
hydrate risk management strategy, two key problems have been accurately. Then, Ding et al.23 conducted a series of flow loop
widely investigated: hydrate formation kinetics5−12 and hydrate experiments and made two flow patterns maps, one with
slurry flow properties.13−20 At present, most of these studies
focus on the water or water/oil emulsion system, and very few Received: July 18, 2016
of them involve a gas−liquid two phase flow, which is the most Revised: March 8, 2017
common system in subsea pipeline transportation. So, the Accepted: March 14, 2017
interaction effect between hydrate formation and the gas− Published: March 14, 2017

© 2017 American Chemical Society 4173 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

hydrate particles and the other one without hydrate particles. systems, hydrate would first form at the oil−water interface as a
Through the comparison between these two flow pattern maps, hydrate shell. Then, the hydrate plug formed mainly due to the
they confirmed that the hydrate formation would cause three shell growth and the decrease of hydrate transportability, which
main changes of the flow pattern map: (i) the area of the was dependent on the pump speed and water cut. In addition,
stratified smooth flow decreases, and the stratified smooth flow they simplified and improved the hydrate formation model and
will transform into the slug flow or the stratified wave flow at found that the effective diffusivity and the hydrate−oil slip ratio
smaller gas/liquid velocities; (ii) the boundary of the annular were the most sensitive parameters with respect to the plugging
flow moves a little to the left, meaning that the slug flow and tendency.32 For water dominant systems, Joshi et al.33 divided
the stratified wave flow will transform into the annular flow at the hydrate plug formation process into three stages based on
smaller gas velocities; (iii) the boundary between the slug flow their experimental results: stage I consists of constant pump
and the bubble flow slightly moves down, and it is easier for the ΔP, stage II consists of a sharp increase in the pump ΔP, and
slug flow to transform into the bubble flow. Besides, in the stage III consists of large fluctuations in the pump ΔP. Then,
model research aspect, Zerpa et al.24 established a hydro- they pointed out that the hydrate plug formation in water
dynamic slug model that considered the gas−liquid-hydrate dominant systems was a consequence of the increase of hydrate
flow in the gas−water system. Their results indicated that the concentration, which would further lead to the formation of
hydrate formation would induce a flow regime transition from hydrate bed and wall deposit. The mechanism of hydrate plug
the stratified flow to the slug flow, which was consistent with formation in gas dominant systems has been studied by Rao et
the experimental observation of Joshi.21 Then, Hegde et al.25 al.34 They found that, in gas dominant systems, hydrate would
used the model established by Zerpa et al.24 to predict the deposit on the pipe wall, starting from nucleation to dendritic
effects of hydrates on the slug characteristics, such as the slug growth to annealing/hardening of the deposit. Also, they
length distribution, number of slugs, and slug frequency. Their proposed a model to predict the hydrate deposition process,
results showed that the liquid-hydrate slip, hydrate volume and results indicated that the hydrate thickness and the distance
fraction, and hydrate aggregation affected the slug character- of plug formation length were significantly affected by the water
istics significantly. Then, Rao et al.26 used a hydrodynamic slug saturation and fluids velocity. Hydrate plugging mechanisms in
model coupled with a transient hydrate formation model to these three systems are briefly shown in Figure 1.
simulate the gas−liquid flow in subsea pipeline. This model can
predict the flow regime transition among the stratified flow, the
stratified wave flow, the slug flow, and the bubble flow, both
with and without hydrate particles.
The above studies have uncovered the mechanisms of how
hydrate formation influences the multiphase flow properties,
and effective models have been proposed to predict the flow
property variation. However, for the influence of different
multiphase flow factors on hydrate formation kinetics, there are
very few relevant studies.
Lv et al.22 studied the influence of gas/liquid flow rates on
gas-slurry flow pressure drop and found that the influence of
liquid superficial velocity on the pressure drop was more
obvious than that of the gas superficial velocity in the stratified
flow; but they did not clarify the influence of gas/liquid flow Figure 1. Schematic diagram of the hydrate plugging mechanism in
rates on hydrate formation kinetics, which is very important in different systems.
modifying the hydrate growth model in multiphase flow
systems. Lorenzo et al.27,28 investigated the hydrate formation
process in annular flow systems; their results confirmed that the Based on the above studies, mechanisms of how hydrate
hydrate growth rate in a gas dominant system was significantly formation affect the multiphase flow properties have been well
larger than that in the water or oil dominant systems. This addressed; however, in turn, the influence of different
indicates that changing the gas/liquid flow rates (or gas/liquid multiphase flow parameters on hydrate formation kinetics is
volume fractions) can influence the hydrate formation rate. In still unclear, especially the influence of different flow patterns
addition, they also pointed out that, in the annular flow, the on hydrate agglomeration and deposition properties. In the
plugging mechanism was dependent on the supercooling present work, a series of experiments were conducted using a
degree of the experimental system. Then, Cassar et al.29 high pressure flow loop. Hydrate agglomeration and deposition
conducted hydrate formation experiments in both the annular properties were studied, and the plugging mechanisms in
flow system and the stratified flow system. They found that in different flow patterns were proposed.
both systems the line blockage was reached after three steps:
(1) rapid hydrate formation and growth, (2) hydrate formation 2. EXPERIMENTAL SECTION
rate slowdown, and (3) the increase of the formation rate, and 2.1. High Pressure Hydrate Flow Loop. The experiments
they also found that the gas−water flow pattern affected the in this work were conducted using a high pressure flow loop,
hydrate formation rate and plugging time apparently. which was constructed by the State Key Laboratory of Pipeline
The hydrate plug formation mechanism is different in Safety in China University of Petroleum (Beijing). The loop
different flow systems, which has been studied by many consists of a centrifugal pump, a gas compressor, four test
researchers. Davies and Boxall et al.30 improved the hydrate sections, a data acquisition system, and several data sensors.
formation and plugging mechanism proposed by Turner31 for The test section is 30 m long in total, and the internal diameter
oil dominant systems. They pointed out that, in oil dominant is 2.54 cm. It is made from carbon steel, and the design pressure
4174 DOI: 10.1021/acs.iecr.6b02717
Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

Figure 2. Top - schematic diagram of the high pressure hydrate flow loop; bottom -photograph of the flow loop test section.

is 15 MPa. The working temperature of the flow loop ranges equipped with two flow meters, one for the liquid flow rate and
from −20 to 100 °C, which is controlled by four Julabo water the other one for the gas flow rate. On the test section, a
baths with a precision of 0.01 °C. Besides, the loop is equipped focused beam reflectance measurement (FBRM) probe and a
with 5 pressure sensors and 8 temperature sensors, with the particle video microscope (PVM) probe are quipped, which can
precision of 0.01 bar and 0.1 °C, respectively. It is also help to study the size and behaviors of hydrate particles from a
4175 DOI: 10.1021/acs.iecr.6b02717
Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

Table 1. List of Experiments

exp. no. flow pattern Qg (kg/h) Ql (kg/h) P (MPa) exp. no. flow pattern Qg (kg/h) Ql (kg/h) P (MPa)
1 stratified 50 500 5 11 slug 100 1100 5
2 stratified 80 500 5 12 slug 100 1100 5
3 stratified 100 700 5 13 slug 100 1100 5
4 stratified 100 500 6 14 slug 100 1100 6
5 bubble 50 1200 5 15 slug 100 1250 6
6 bubble 50 1750 5 16 slug 150 1100 5
7 bubble 60 1100 5 17 annular 180 450 5
8 bubble 70 1800 6 18 annular 260 400 5
9 slug 120 1100 5 19 annular 260 500 6
10 slug 100 1100 5

microscopic view. In front of the gas/liquid mixer there is a

density meter, which can measure the changes of the slurry
density. In addition, there are four glass windows equipped on
the test section, from which we can directly observe the flow
patterns and flow conditions in the pipeline. A photograph of
the test section and a process flow diagram are shown in Figure
2.
2.2. Materials and Procedures. The materials used in the
experiments include civil natural gas from Shanjing Natural Gas
Pipeline in China, deionized water, −20# diesel oil, and AAs.
For the composition of the gas and diesel oil, please refer to our
previous work.23 The AA is not an industrial AA, and it is
extracted from a saponins plant that was developed by the
Chemical Engineering Department in China University of
Petroleum-Beijing.35
The experiments were carried out with 10% water cut and
1% AAs dosage. The water cut was defined as the ratio of the
water volume to the whole liquid volume. The AA dosage is Figure 3. Changes of chord length distribution in Exp. 15.
defined as the volume fraction of AA additive in the water
phase. The setting temperature of the water bath was 0 °C. For
each experiment, the flow pattern was confirmed through the The three lines in Figure 3 stand for the particles/droplets
observation from the glass window. The experiment list is chord length distributions at three different time points in Exp.
shown in Table 1. 15: just before hydrate formation (black line), 10 min after
The experimental procedure is briefly introduced as follows: hydrate formation (red line), and the final state of the
(i) Vacuum the loop to −1 bar using a vacuum pump to experiment (blue line). We can see that there is an intersection
eliminate the influence of air. (ii) Load the deionized water and point of the three lines at about 40 μm, and we define this point
diesel oil at a specified water cut. (iii) Inject natural gas into the as the critical chord length of hydrate agglomeration in Exp. 15.
loop to reach the experimental pressure. (iv) Circulate the After hydrate formation onset, due to the agglomeration
liquid and gas by the pump and compressor, respectively. The between hydrate particles and water droplets, the number of
gas/liquid flow rates can be adjusted to form different gas− particles/droplets smaller than Cc decreased apparently, while
liquid flow patterns, in which the hydrates will form in the next the number of the particles/droplets larger than Cc increased
step. (v) When the monitored flow parameters reach a stable rapidly. Thus, we use the changes of the percentage of hydrate
state, cool down the loop to form hydrate crystals. In this particles larger than Cc to estimate the hydrate agglomeration
period, both the flow parameters and the hydrate particles degree
behaviors are recorded. (vi) When all the collected data keep fa = max φC c − φC0c
stable, increase the temperature to decompose the hydrates.
Then, start another set of experiments in another flow pattern (1)where fa is the factor of hydrate agglomeration degree, φCc is
according to the above steps. the number percentage of hydrate particles/water droplets
2.3. Methods of Quantitative Estimation for the larger than Cc, and φ0Cc is the number percentage of water
Hydrate Agglomeration and Deposition Degree. The droplets larger than Cc before hydrate formation.
main objective of this work is to study the hydrate However, this method is inapplicable for the annular flow
agglomeration and deposition properties in different flow because no Cc occurred in annular flow experiments. FBRM
patterns and then analyze the plugging mechanisms in each results of Exp. 19 in the annular flow are shown in Figure 4.
flow pattern. In this section, several methods were proposed to After hydrate formation onset, all particles decreased in
estimate the factor of hydrate agglomeration degree, fa, and the number, and there was no intersection point of the three
factor of hydrate deposition degree fd. lines in Figure 4. This was because in the annular flow hydrate
The first method to estimate fa is based on the critical chord formed mainly on the pipe wall surface in the form of a hydrate
length of hydrate agglomeration, Cc. Here we use the FBRM layer, so hydrate agglomeration in the bulk phase was not
results of Exp. 15 as an example, which is shown in Figure 3. obvious. Therefore, another method is introduced to estimate
4176 DOI: 10.1021/acs.iecr.6b02717
Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

decrease gradually. This decrease of the large particles was not

caused by the agglomerates breaking, because the total number
of hydrate particles also decreased slightly instead of increasing
at this period. Therefore, this was likely caused by the hydrate
deposition on the pipe wall surface. In addition, it was worth
noting that the slurry density, ρs, also decreased apparently after
the hydrate formation. As we know, hydrate density is larger
than that of the oil phase (about 740 kg/m3), and the
unconverted water which is encapsulated in the hydrate
agglomerates may also deposit together with hydrate particles.
Therefore, hydrate deposition can lead to a significant
reduction of the slurry density. Thus, based on these two
phenomena, we concluded that hydrate deposition occurred in
this experiment. Here we use the level of the slurry density
reduction to estimate the factor of hydrate deposition degree,
fd, which is calculated by
Figure 4. Changes of chord length distribution in Exp. 19.
(ρs − initial − ρs − final ) φH0
fd =
the hydrate agglomeration degree which is based on the square- (ρs − initial − ρoil ) φH (3)
weighted mean chord length of hydrate particles, Cm. Changes
of the Cm of in Exp. 19 are shown in Figure 5. We can see that where ρs−initial is the slurry density just before hydrate formation
onset (kg/m3), ρs−final is the slurry density at the final state of
the experiment (kg/m3), ρoil is the density of the oil phase (kg/
m3), φH is the total volume fraction of hydrates formed, and φ0H
is the hydrate volume fraction under the condition that the
added water is totally converted into hydrates. The second item
at the right side of the equation is to eliminate the influence of
different hydrate formation amount on the hydrate deposition
process. The hydrate formation amount was obtained through
the gas consumption amount, which was calculated by
PV
1 PV
ng = − 2
z1RT1 z 2RT2 (4)
where ng is the mole number of gas consumption (mol), P1 is
the pressure before hydrate formation (Pa), P2 is the pressure
after hydrate complete formation (Pa), V is the gas volume in
the separator (m3), z is the compressibility factor in
experimental pressure, R is the gas constant (J/mol/K), T1 is
the temperature before hydrate formation (K), and T1 is the
Figure 5. Changes of the square-weighted mean chord length in Exp. temperature after hydrate complete formation (K).
19. Based on the gas consumption amount, the hydrate volume
fraction can be calculated by
the Cm increased rapidly after hydrate formation onset in the ng Mg + N *ng M w
annular flow. So we can use the increase of Cm to estimate the φH =
hydrate agglomeration degree, which can be calculated by ρH VL (5)

max Cm − Cm0 where φH is the hydrate volume fraction, Mg is the hydrate

f a′ = molar mass (g/mol), N is the hydration number (5.85 for
Cm0 (2) natural gas), Mw is the water molar mass (g/mol), ρH is the
where fa′ is the factor of hydrate agglomeration degree hydrate density (kg/m3), and VL is the total volume of the
calculated by Cm, and C0m is the square-weighted mean chord liquid phase (m3).
length before hydrate formation onset.
Both of the above two methods can be used to estimate the 3. RESULTS AND DISCUSSION
hydrate agglomeration degree, and the results of these two During the experiments, the flow patterns were confirmed by
methods are compared in section 3.4. the visual observation through the glass windows, and the
The method to estimate hydrate deposition degree is based following four flow patterns were studied in this work: the
on the changes of slurry density ρs. Here we also use the results stratified flow, the bubble flow, the slug flow, and the annular
of Exp. 15 as an example, which is shown in Figure 6. We can flow. In this section, typical results of the experiments in
see from Figure 6, after the hydrate formation onset the total different flow patterns were presented and analyzed. Then,
number of hydrate particles/water droplets, Np, decreased based on the analysis of the typical results, plugging
apparently, and the number of particles larger than 100 μm, Np mechanisms in different flow patterns were discussed. In
> 100, increased rapidly, which indicated that the hydrate addition, degrees of hydrate agglomeration and hydrate
agglomeration occurred at this time. Then, Np >100 began to deposition in each flow pattern were compared.
4177 DOI: 10.1021/acs.iecr.6b02717
Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

Figure 6. Changes of the slurry density and particle number in Exp. 15.

Figure 7. Results of Exp. 4 in stratified flow condition.

3.1. Results in Stratified Flow. Exp. 1−4 were carried out liquid phase as we can see the total number of particles
in stratified flow conditions, and the results of these four increased during this period; and finally, the loop was blocked,
experiments were similar. Here we use the results of Exp. 4 as which is shown in Figure 8 (b). We can notice that the slurry
an example to show the typical results in the stratified flow. density changed very little after hydrate formation, indicating
As shown in Figure 7, the relative DP is the pressure drop of
the flow loop divided by the liquid flow rate. When hydrate
began to form at about 1.7 h, the total number of hydrate
particles/water droplets decreased rapidly, indicating the
hydrate agglomeration occurred at this time. Due to the
hydrate formation and violent agglomeration, the liquid flow
rate decreased and the relative DP increased rapidly. After this
period, both the relative DP and the liquid flow rate kept
constant for about half an hour. Then again the liquid flow rate
began to decrease and the relative DP began to increase. This Figure 8. (a) Hydrate formation and (b) hydrate plugging in the
was caused by the hydrate growth and accumulation in the stratified flow.

4178 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

that the hydrate deposition degree in the stratified flow was factors such as the breaking of hydrate-coated bubbles, the
very small. In addition, the total number of hydrate particles in breaking of hydrate agglomerates, or the hydrate continuous
the liquid phase increased continuously after the initial growth in the liquid phase. Figure 11 shows the changes of the
agglomeration, which also meant the hydrate particles tend to
grow in the liquid phase instead of depositing on the pipe wall
surface. So the plugging in this experiment was mainly caused
by the continuous growth and accumulation of hydrates in the
liquid phase, as shown in Figure 8 (b).
Based on the above analysis and the recorded picture, a plug
formation mechanism in the stratified flow is proposed, as
shown in Figure 9: (i) The system stays at the gas−liquid

Figure 9. Schematic diagram of plugging mechanism in the stratified

flow (adapted from the diagram proposed by Turner31).

stratified flow with the water dispersed in the oil phase as water
droplets; (ii) When the system runs into the hydrate stable Figure 11. Changes of mean chord length and number of particles in
Exp. 5.
region, hydrate nucleation onset occurred; (iii) Hydrate
particles and water droplets begin to agglomerate with each
other; (iv) After the rapid agglomeration, the agglomerates square-weighted mean chord length of the particles, which
grow continuously; (v) Hydrates accumulate in the liquid phase shows the similar trend with the change of total number of
and then bedding on the pipe wall, and then the pipeline is particles. This indicates that the increase of particles number is
blocked. not caused by the agglomerate breaking or bubble breaking,
3.2. Results in Bubble Flow. Exp. 5−8 were carried out since this would cause reduction of the mean chord length. So
with very small gas flow rates in order to form bubble flow in this stage, hydrate grew continuously in the liquid phase.
conditions. Results in these experiments are similar, and here Then, the total number of particles began to decrease, along
we use the results of Exp. 5 as an example, which is shown in with the slurry density and the liquid flow rate. From the
Figure 10. reduction of the slurry density and particles number, we
As shown in Figure 10, after the hydrate formation onset at deduced that this was caused by the hydrate deposition on pipe
about 1.7 h, the total number of particles/droplets/bubbles wall surface. Then, the deposition process ceased, and the
decreased rapidly, indicating that hydrate agglomeration system kept a stable state. No plug formed in this experiment.
occurred at this period. Due to the hydrate formation and Based on the above results, the hydrate formation and slurry
agglomeration, the liquid flow rate decreased gradually and the flow process in the bubble flow is proposed, as shown in Figure
relative DP increased rapidly. Then the total number of 12: (i) Water disperses in the oil phase as water droplets, and
particles began to increase, which could be caused by several the system maintains at a stable bubble flow; (ii) When the

Figure 10. Results of Exp. 5 in bubble flow condition.

4179 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

the relative DP. However, this phenomenon did not occur in

every slug flow experiment and may not be regarded as the
typical features of the slug flow. Then the system kept a stable
state until the end, and no plug formed at last.
Based on the above results, the hydrate formation and slurry
Figure 12. Schematic diagram of hydrate particles behaviors in the
flow process in the slug flow is proposed as shown in Figure 15:
bubble flow.
(i) The system maintains at the stable slug flow with water
droplets dispersed in the liquid phase; (ii) When the system
system runs into a hydrate stable region, hydrates begin to form runs into a hydrate stable region, hydrates begin to nucleate on
on the water/oil interface; (iii) Hydrate particles and water the water/oil interface; (iii) Hydrates and water droplets begin
droplets begin to agglomerate with each other, forming large to agglomerate in the liquid phase and deposit on the pipe wall
agglomerates; (iv) Hydrates and agglomerates grow in the surface at the same time; (iv) Some of the agglomerates are
liquid phase; (v) Hydrates and agglomerates deposit on the broken up by the flow shear force (may not occur); (v) The
pipe wall surface, and then the system maintains a stable flow system maintains a stable flow state, and no plug formed in the
condition. flow loop.
3.3. Results in Slug Flow Condition. Exp. 9−16 were 3.4. Results in Annular Flow Condition. Exp. 17−19
carried out in slug flow conditions. Results of Exp. 10 are used were carried out in annular flow conditions, in which the liquid
here to give an introduction of the typical results in the slug phase flows close to the pipe wall as annulus and the gas phase
flow, which is shown in Figure 13. flows continuously in the center of the pipe. Results of Exp. 17
Results of the slug flow are similar to that of the bubble flow. are shown in Figure 16.
When hydrates began to form, the total number of particles/ As shown in Figure 16, the hydrate formation onset occurred
droplets decreased apparently in a very short time. This was at about 2.5 h. Then the slurry density and the number of
caused by both the hydrate agglomeration in the liquid phase particles decreased immediately. This was not because of the
and the hydrate deposition on the pipe wall surface. Because we hydrate deposition on the pipe wall, but because hydrates
can see from Figure 13 and Figure 14, at this time, the slurry formed directly on the pipe wall surface. A thick hydrate layer
density decreased, and also the square-weighted mean chord was observed on the inner surface of the pipe wall at this time,
length increased rapidly. This indicated that the hydrate as shown in Figure 17(a). Because of the formation of this
agglomeration and deposition occurred at the same time, hydrate layer, the liquid flow rate decreased and the relative DP
which caused a large decrease of the liquid flow rate and a large increased. Then both the slurry density and the number of
increase of the relative DP. The deposition process in the slug particles began to increase, and at the same time we observed
flow occurred much earlier than that in the bubble flow and the that the formed hydrate layer began to slough, as shown in
stratified flow. This was because the slug flow had large flow Figure 17(b). Then the liquid flow rate decreased to 0, and the
fluctuations, and thus the hydrate particles could be carried and loop was blocked in a very short time. Based on the changes of
contact with the pipe wall at high frequency. Therefore, the slurry density and particles number and the recorded
deposition occurred immediately after hydrate formation onset pictures in Figure 17, we deduced that the plugging was
in the slug flow. This process lasted about 40 min, and then the because the sloughed-off hydrate fragments accumulated in the
deposition process ceased as the slurry density almost kept pipe and were stuck somewhere in the flow loop like the elbows
constant. We can notice that, during this period, the mean or valves.
chord length decreased a little, and the total number of particles The plugging mechanism in the annular flow is shown in
increased slightly. This indicated that some of the hydrate Figure 18: (i) The system maintains at the annular flow before
agglomerates broke up, which also led to a slight reduction of hydrate formation. Some of the water disperses in the oil phase

Figure 13. Results of Exp. 10 in the slug flow.

4180 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

Figure 14. Changes of mean chord length and number of particles in Exp. 5.

Figure 15. Schematic diagram of hydrate particles behaviors in the slug

flow.

as water droplets, and some distributes as water film covering Figure 17. (a) Hydrate layer formed on the glass window and (b)
the pipe wall; (ii) Hydrates begin to form on the pipe wall or hydrate film sloughing.
on the water/oil interface, forming a thick hydrate layer
covering the pipe wall; (iii) The thick hydrate layer begins to including the factor of hydrate agglomeration degree fa and f′a,
slough due to the intense flow shear force; (iv) The sloughed the factor of hydrate deposition degree fd, and the hydrate
hydrate fragments accumulate at some uneven section and volume fraction φH. Detailed results are listed in Table 2.
block the flow section. The volume fraction of hydrates formed in each flow pattern
3.5. Comparison of the Results in Each Flow Pattern. is shown in Figure 19. We can see that the hydrate volume
In section 2.3, several methods were proposed to estimate the fraction (or hydrate formation amount) in the bubble flow and
hydrate agglomeration degree and deposition degree. In this the slug flow has the maximum value, both of which are about
section, results in different flow patterns are compared, 4.4%; but the error range in the slug flow is larger. The hydrate

Figure 16. Results of Exp. 17 in annular flow condition.

4181 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

Figure 18. Schematic diagram of the plugging mechanism in the annular flow (adapted from the diagram proposed by Sum et al.36).

Table 2. Detailed Results for Each Experiment

exp. no. flow pattern φH fa fa′ fd exp. no. flow pattern φH fa fa′ fd
1 stratified 0.034 0.3 0.86 0.29 11 slug 0.037 0.58 1.37 1.85
2 stratified 0.036 0.46 1.31 0.30 12 slug 0.051 0.62 1.37 1.26
3 stratified 0.046 0.56 1.79 0.31 13 slug 0.046 0.58 1.30 1.03
4 stratified 0.036 0.33 0.98 0.36 14 slug 0.048 0.59 1.54 1.10
5 bubble 0.043 0.4 1.09 0.81 15 slug 0.048 0.64 1.77 1.40
6 bubble 0.044 0.15 0.38 1.47 16 slug 0.038 0.72 2.86 2.21
7 bubble 0.046 0.34 1.16 0.53 17 annular 0.014 0.49 1.90
8 bubble 0.043 0.19 0.48 1.55 18 annular 0.018 0.27 1.92
9 slug 0.039 0.66 2.89 2.33 19 annular 0.016 0.63 2.33
10 slug 0.047 0.75 2.79 2.18

The factors of hydrate agglomeration degree and deposition

degree in each flow pattern are shown in Figure 20. fa is

Figure 19. Hydrate volume fraction in each flow pattern.

volume fraction in stratified flow is about 3.8%, and the hydrate Figure 20. Factors in different flow patterns.
volume fraction in the annular flow is only about 1.6%. We
should mention here that all the experiments in stratified flow
and annular flow conditions were blocked at last. The plugging calculated based on the critical chord length, and f′a is calculated
process in the annular flow was very rapid, while the plugging based on the square-weighted mean chord length. We can see
process in the stratified flow occurred gradually. Because of the that, for each flow pattern, the ratio of fa′/fa is almost a constant
plugging, the length of the hydrate growth period is different in of 3, which demonstrates that both of the above two methods
different flow patterns. Thus, the blockage is likely to be the are valid for estimating the agglomeration degree. The results
reason for the difference of the hydrate formation amount. As show that these two factors have the same change tendency: the
we know, the hydrate formation amount is mainly affected by slug flow > the stratified flow > the bubble flow > the annular
the experimental pressure, temperature, and the water cut. So as flow. This indicates that hydrates in the slug flow have the
long as the flow system keeps a good flow stability, the hydrate largest agglomeration degree, which may be due to the unstable
formation amount in each flow pattern should be very close to flow condition. Because the slug flow has violent flow
each other. fluctuations, hydrate particles in the slug flow can contact and
4182 DOI: 10.1021/acs.iecr.6b02717
Ind. Eng. Chem. Res. 2017, 56, 4173−4184
Industrial & Engineering Chemistry Research Article

collide with each other more frequently. Thus, the agglomer- (3) Zarinabadi, S.; Samimi, A. Problems of hydrate formation in oil
ation degree is higher in the slug flow. In the annular flow, and gas pipes deals. Aus. J. Basic Appl. Sci. 2011, 5, 741.
however, hydrates tend to form and grow directly on the pipe (4) Creek, J. L. Efficient Hydrate Plug Prevention. Energy Fuels 2012,
wall surface (Figure 17), so the agglomeration in the liquid 26, 4112.
phase is not obvious. Therefore, the agglomeration degree in (5) Dai, S.; Lee, J. Y.; Santamarina, J. C. Hydrate nucleation in
quiescent and dynamic conditions. Fluid Phase Equilib. 2014, 378, 107.
the annular flow is the smallest.
(6) Fandiño, O.; Ruffine, L. Methane hydrate nucleation and growth
Also, Figure 20 shows that the change tendency of the from the bulk phase: Further insights into their mechanisms. Fuel
hydrate deposition degree is the annular flow > the slug flow > 2014, 117, 442.
the bubble flow > the stratified flow. As discussed above, in (7) Veluswamy, H. P.; Linga, P. Macroscopic kinetics of hydrate
annular flow hydrates tend to form and grow directly on the formation of mixed hydrates of hydrogen/tetrahydrofuran for
pipe wall surface, which can lead to a higher deposition degree. hydrogen storage. Int. J. Hydrogen Energy 2013, 38, 4587.
For the stratified flow, because hydrate deposition mainly (8) Kroeger, K.; Plaza-Faverola, A.; Barnes, P.; Pecher, I. Thermal
occurs at the liquid−solid interphase, the liquid−solid interface evolution of the New Zealand Hikurangi subduction margin: Impact
in the stratified flow is the smallest compared with other flow on natural gas generation and methane hydrate formation−A model
patterns. That is likely to be the reason why the stratified flow study. Mar. Pet. Geol. 2015, 63, 97.
has the smallest hydrate deposition degree. (9) Shi, B.-H.; Gong, J.; Sun, C.-Y.; Zhao, J.-K.; Ding, Y.; Chen, G.-J.
An inward and outward natural gas hydrates growth shell model
considering intrinsic kinetics, mass and heat transfer. Chem. Eng. J.
4. CONCLUSIONS
2011, 171, 1308.
Experiments were carried out using a high pressure flow loop to (10) Lv, X.; Gong, J.; Li, W.; Shi, B.; Yu, D.; Wu, H. Experimental
investigate the hydrate behaviors and the slurry plugging study on natural-gas-hydrate-slurry flow. SPE J. 2014, 19, 206.
mechanism in different flow patterns. Based on the changes of (11) Daraboina, N.; Pachitsas, S.; von Solms, N. Natural gas hydrate
slurry density and the particle chord length distribution, new formation and inhibition in gas/crude oil/aqueous systems. Fuel 2015,
methods were proposed to estimate the degree of hydrate 148, 186.
agglomeration and deposition in different flow patterns. Results (12) Najibi, H.; Shayegan, M. M.; Heidary, H. Experimental
showed that the agglomeration degree in order from high to investigation of methane hydrate formation in the presence of copper
oxide nanoparticles and SDS. J. Nat. Gas Sci. Eng. 2015, 23, 315.
low is the slug flow > the stratified flow > the bubble flow > the
(13) Sohn, Y. h.; Kim, J.; Shin, K.; Chang, D.; Seo, Y.; Aman, Z. M.;
annular flow; the deposition degree in order from high to low is May, E. F. Hydrate plug formation risk with varying watercut and
the annular flow > the slug flow > the bubble flow > the inhibitor concentrations. Chem. Eng. Sci. 2015, 126, 711.
stratified flow. In addition, typical results of the experiments in (14) Peng, B.-Z.; Chen, J.; Sun, C.-Y.; Dandekar, A.; Guo, S.-H.; Liu,
different flow pattern conditions were presented. It was found B.; Mu, L.; Yang, L.-Y.; Li, W.-Z.; Chen, G.-J. Flow characteristics and
that the slurry flow in the stratified flow and annular flow morphology of hydrate slurry formed from (natural gas+ diesel oil/
conditions was easily blocked. The plugging in the stratified condensate oil+ water) system containing anti-agglomerant. Chem.
flow was mainly due to the hydrate accumulation and bedding Eng. Sci. 2012, 84, 333.
in the liquid phase, while the plugging in the annular flow was (15) Wang, W.; Fan, S.; Liang, D.; Li, Y. Experimental study on flow
mainly caused by the hydrate layer sloughing. characteristics of tetrahydrofuran hydrate slurry in pipelines. J. Nat.

■
Gas Chem. 2010, 19, 318.
AUTHOR INFORMATION (16) Pauchard, V.; Darbouret, M.; Palermo, T.; Peytavy, J.-L. Gas
hydrate slurry flow in a black oil. Prediction of gas hydrate particles
Corresponding Authors agglomeration and linear pressure drop. Proc. 13th International
*E-mail: ydgj@cup.edu.cn. Conference on Multiphase Production Technology, Edinburgh, UK, 23−15
*E-mail: bh.shi@cup.edu.cn. June 2007.
(17) Sun, M.; Firoozabadi, A. Natural gas hydrate particles in oil-free
ORCID
systems with kinetic inhibition and slurry viscosity reduction. Energy
Jing Gong: 0000-0002-3722-5778 Fuels 2014, 28, 1890.
Notes (18) Sinquin, A.; Palermo, T.; Peysson, Y. Rheological and flow
The authors declare no competing financial interest. properties of gas hydrate suspensions. Oil Gas Sci. Technol. 2004, 59,

■
41.
(19) Andersson, V.; Gudmundsson, J. Flow experiments on
ACKNOWLEDGMENTS concentrated hydrate slurries. 1999 SPE Annual Technical Conference
This work was supported by the National Science Foundation and Exhibition: Production Operations and Engineering - General,
for Young Scientists of China (Grant 51306208), National Houston, TX, 3−6 Oct., 1999; p 39310.2118/56567-MS.
Natural Science Foundation of China (Grant 51274218 & (20) Lv, X.; Shi, B.; Wang, Y.; Tang, Y.; Wang, L.; Gong, J.
51534007), National Science and Technology Major Project Experimental Study on Hydrate Induction Time of Gas-Saturated
(No. 2016ZX05028004-001), and Science Foundation of China Water-in-Oil Emulsion using a High-Pressure Flow Loop. Oil Gas Sci.
University of Petroleum-Beijing (No. 2462014YJRC006, No. Technol. 2015, 70, 1111.
2462015YQ0404, and No. C201602), which are gratefully (21) Joshi, S. V. Experimental investigation and modeling of gas
hydrate formation in high water cut producing oil pipelines. Ph.D.
acknowledged.

■
Dissertation, Colorado School of Mines, Golden, CO, 2012.
(22) Lv, X.; Shi, B.; Wang, Y.; Gong, J. Study on Gas Hydrate
REFERENCES Formation and Hydrate Slurry Flow in a Multiphase Transportation
(1) Sloan, E.; Koh, C.; Sum, A.; Ballard, A.; Shoup, G.; McMullen, System. Energy Fuels 2013, 27, 7294.
N.; Creek, J.; Palermo, T. Hydrates: state of the art inside and outside (23) Ding, L.; Bohui, S.; Xiaofang, L.; Yang, L.; Haihao, W.; Wei, W.;
flowlines. J. JPT, J. Pet. Technol. 2009, 61, 89. Jing, G. Investigation of natural gas hydrate slurry flow properties and
(2) Sloan, E. D.; Koh, C. A.; Sum, A. Natural gas hydrates in flow flow patterns using a high pressure flow loop. Chem. Eng. Sci. 2016,
assurance; Gulf Professional Publishing: Oxford, UK, 2010. 146, 199.

4183 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184
 Industrial & Engineering Chemistry Research Article

(24) Zerpa, L. E.; Rao, I.; Aman, Z. M.; Danielson, T. J.; Koh, C. A.;
Sloan, E. D.; Sum, A. K. Multiphase flow modeling of gas hydrates with
a simple hydrodynamic slug flow model. Chem. Eng. Sci. 2013, 99, 298.
(25) Hegde, G. A.; Sum, A. K.; Danielson, T. J. Multiphase Flow
Modeling for Gas Hydrates in Flow Assurance. Offshore Technology
Conference, 04−07 May, 2015. Houston, Texas, USA, OTC-
25729.10.4043/25729-MS
(26) Rao, I.; Sum, A. K.; Koh, C. A.; Sloan, E. D.; Zerpa, L. E.
Multiphase Flow Modeling of Gas-Water-Hydrate Systems. Offshore
Technology Conference, Houston, USA, 2013. OTC-24099.
(27) Di Lorenzo, M.; Aman, Z. M.; Sanchez Soto, G.; Johns, M.;
Kozielski, K. A.; May, E. F. Hydrate formation in gas-dominant
systems using a single-pass flowloop. Energy Fuels 2014, 28, 3043.
(28) Di Lorenzo, M.; Aman, Z. M.; Kozielski, K.; Norris, B. W.;
Johns, M. L.; May, E. F. Underinhibited Hydrate Formation and
Transport Investigated Using a Single-Pass Gas-Dominant Flowloop.
Energy Fuels 2014, 28, 7274.
(29) Cassar, C.; Sinquin, A.; Teixeira, A.; Glenat, P.; Leininger, J. In
Hydrate formation in gas dominant systems: influence of the flow
pattern. 12th Offshore Mediterranean Conference and Exhibition,
Ravenna, Italy, 2015.
(30) Boxall, J. A.; Davies, S. R.; Nicholas, J. W. et al. Hydrate
blockage potential in an oil-dominated system studied using a four
inch flow loop. Proceedings of the Fourth International Conference of Gas
Hydrates. Vancouver, British Columbia, Canada, July 6−10, 2008.
(31) Turner, D. J., Clathrate hydrate formation in water-in-oil
dispersions. Ph.D. Thesis, Colorado School of Mines, Golden, CO,
2006.
(32) Davies, S. R.; Boxall, J. A.; Dieker, L. E.; et al. Predicting hydrate
plug formation in oil-dominated flowlines. J. Pet. Sci. Eng. 2010, 72,
302.
(33) Joshi, S. V.; Grasso, G. A.; Lafond, P. G.; et al. Experimental
flowloop investigations of gas hydrate formation in high water cut
systems. Chem. Eng. Sci. 2013, 97, 198.
(34) Rao, I.; Sloan, E. D.; Koh, C. A.; Sum, A. K. Laboratory
experiments and modeling for hydrate formation and deposition from
water saturated gas systems. Proceedings of the 7th International
Conference on Gas Hydrates; Edinburgh, U.K., July 17−21, 2011.
(35) Chen, J.; Wang, Y.-F.; Sun, C.-Y.; Li, F.-G.; Ren, N.; Jia, M.-L.;
Yan, K.-L.; Lv, Y.-N.; Liu, B.; Chen, G.-J. Evaluation of Gas Hydrate
Anti-agglomerant Based on Laser Measurement. Energy Fuels 2015, 29,
122.
(36) Sum, A. K.; Koh, C. A.; Sloan, E. D. Developing a
comprehensive understanding and model of hydrate in multiphase
flow: from laboratory measurements to field applications[J]. Energy
Fuels 2012, 26, 4046.

4184 DOI: 10.1021/acs.iecr.6b02717

Ind. Eng. Chem. Res. 2017, 56, 4173−4184

View publication stats