Heat Transfer and Multiphase Flow with Hydrate
Formation in Subsea Pipelines
A. Odukoya *, G. F. Naterer
Faculty of Engineering and Applied Science
Memorial University, St. Johns, Newfoundland
St. Johns, Newfoundland, Canada
* Email: aodukoya@mun.ca
Abstract This study presents a new semi-analytical model
to predict hydrate formation in subsea pipelines. The numerical
formulation is based on a non-dimensional analysis, allowing the
model to be extended to various pipeline lengths and
environmental conditions. The effects of change in heat transfer
ratio (HTR), and pipe diameter on hydrate formation are
reported in this paper. The results indicate that higher HTR
between the internal and external fluids reduces the formation of
hydrates which may lead to a blockage in the pipeline. The
numerical results are validated against experimental results for
R134a hydrates. The pipes with smaller diameters are found to
reduce the possibility of hydrate formation at a constant pipeline
pressure. The results show that at temperatures below -10oC,
changing thermophysical properties have limited impact on the
rate of hydrate formation in the pipe.
Keywords Hydrates; Pipelines; Multiphase flow; Heat
transfer; Phase change
I. INTRODUCTION
The exploration of oil and gas in deep waters and harsh
environments has motivated research on multiphase flows in
subsea pipelines. This requires detailed understanding of the
heat and mass transfer of the oil / gas / water phases in the
pipelines. One of the technical challenges remains predicting
and preventing the formation of hydrates in the pipelines to
ensure the flow assurance in pipelines. The environment
temperature could be as low as -20oC, which falls within
hydrate forming temperatures. Hydrates are formed when the
temperature falls below 303K at ambient pressure, and the
specific structure formed is dependent on the type of gases
flowing through the pipeline. Hydrates act like solid particles
in a flow stream and if not controlled could impede the flow in
the pipeline. Thermodynamic inhibitors (such as MeOH) are
commonly used to prevent hydrate formation in crude oil
pipelines. Other methods of reducing the hydrate formation
include the use of kinetic inhibitors, low dosage hydrate
inhibitors and anti-agglomerates. Heated pipelines are also
used to prevent hydrate formation in short pipeline networks.
The flow of hydrates in a pipeline is a multiphase flow heat
transfer problem. The nucleation of the hydrate is the first step
in the hydrate formation. The nucleation process has been
extensively studied in past literature [1-4]. The hydrate starts
to grow as soon as nucleation is sustained. Some of the
978-1-4799-4918-2/14/$31.00 2014 IEEE
established models for predicting nucleation of hydrates
include Englezos-Kalogerakis-Dhalabhain-Bishno [5],
Skovborg-Rasmussen [6], Herri-Pic-Gruy-Cournil [7], and
Guanendran-Amin [8] models. These studies have investigated
the formation of hydrates taking into consideration the kinetics
of the hydrate formation. But few studies have analyzed the
combined effect of heat and mass transfer on hydrate
formation. Kwon et al. [9] developed a numerical model for a
four-phase fluid hydrodynamic model for oil, water, gas, and
hydrates. The numerical model was solved using a fully
implicit finite difference analysis. Nemeto and Balino [10]
developed a numerical model for risers in oil pipelines to
examine severe slugging in a mixture oil, gas, and water. The
riser model considered in the latter study is a one-dimensional
isothermal flow and the results show nonlinearity in the fluid
flow in the riser. A theoretical analysis of heat and mass
transfer for interacting multiphase mixtures was outlined by
Niessner and Hassanizadeh [11]. The model is well suited for
non-isothermal conditions with heat transfer between phases.
Shagapov et al. [12] developed a quasi stationary model for
gas hydrate dissociation in a horizontal pipe. The model shows
a buildup of the hydrate toward the inlet of the pipe and
indicates that the use of methanol for hydrate dissociation has
limited impact on the rate of hydrate formation beyond a
certain threshold during the coagulation. This paper aims at
developing a new integral formulation to predict the growth of
hydrates in multiphase flows.
Tolubinskiy [13] developed a general solution for transient
non-linear heat and mass transfer problems using an integral
method. The solution is restricted to coupled heat and mass
transfer problems, where the function of the heat source is
defined as a unit pulse from experimental measurements.
Mennig and Ozisik [14] developed an analytical model to
account for time dependent surface conditions during heat
transfer in solidification and melting. Dorstewitz and Mewes
[15] reported experimental and analytical results on R134a
hydrate formation in a pipe. Although heat transfer was
considered in this study, the hydrate layer was modeled as a
uniform layer on the ice wall. Unlike the model of Dorstewitz
and Mewes [15], this paper includes the effects of the
deposition of solid particles on the pipe wall, by considering
both the kinetic energy of the particle and the coagulation of
the hydrate layer over time. Vaze and Banerjee [16] developed
experimental correlations for determining heat transfer
coefficients in two-phase flows. The results indicate that the
heat transfer coefficient could be determined from a known
Nusselt number of the flow regime.
The growth of the hydrate layer is determined by the ability
of the particles to adhere to the pipe walls when formed. The
particles could move along with the fluid flow or decelerate
and adhere to the wall. The irregular shapes formed during
solidification of particles can be defined by imposing boundary
conditions and solving heat and mass balances at the defined
boundaries [17]. Elperin and Fominykh [18] developed a gas
liquid slug model to predict the formation of gas hydrates,
assuming perfect mixing liquid plugs. The result indicates that
the heat and mass transfer of an infinite linear cluster of slugs
is independent of the initial temperature. Jassim et al. [19]
proposed a numerical model to determine the deposition of gas
hydrate particles. A computational fluid dynamics (CFD)
model taking into account the nucleation and growth of gas
hydrates was used to determine the deposition mechanism of
the particle. A new analytical approach is presented in this
paper to determine the growth of the hydrate layer based on the
deposition mechanism.
The heat transfer from the fluid in the pipe to the
environment is often neglected in past models in the literature
[20]. A phase change heat transfer model of Naterer [21] was
incorporated into an ice growth model, to account for the heat
transfer across the wall. This model is used to determine the
effect of the kinetic energy of a droplet on the heat transfer
across the wall. A numerical model simulating the four-phase
(oil, gas, water and hydrate) flow in crude oil pipelines in the
North Sea was proposed by Kwon et al. [9]. The various
regions in the flow were computed using a dew point pressure
line, freezing point line, and hydrate dissociation pressure line.
The results were used to estimate the range of temperature and
pressures that best represent the offshore pipeline temperatures.
The potential hazards of wax formation are a result of
hydrate formation and mitigation techniques were reviewed by
Ayejina et al.[22]. One of the current methods of preventing
wax formation includes heating the pipeline beyond the hydrate
formation temperature. Another established method in the oil
and gas industries is the addition of kinetic inhibitors [23] to
prevent the growth of hydrates. By understanding the
deposition mechanism in relation to heat transfer, the kinetic
inhibitors could be reduced. This paper develops a new
analytical model to analyze the effects of temperature gradient
and pipeline radius on the flow of the hydrate particles. This
paper specifically focuses on R134a hydrates to allow for
comparison with past data of Dorstewitz and Mewes [15].
II. HEAT AND MASS TRANSFER FORMULATION
A force balance on a hydrate particle is used to determine
the velocity of the particle. The forces acting on the particle are
highlighted in Fig. 1. Jassim et al. [19] developed a bouncing
particle model to predict the forces acting on the particle. This
paper extends the bouncing model to include both heat transfer
and kinetic energy of the particle. When hydrate particles
moving at low velocity come into contact with the pipe wall,
the kinetic energy of the particle is reduced and heat is
transferred across the pipe wall along with adjacent particles.
Fig. 1. Schematic of problem description; FD, FL, FG, and FA are the drag,
lift, gravity, and adhesive forces respectively
At high velocity of the particle, it is assumed that the
particle is plastically deformed. This model assumes that
beyond a critical velocity, the particle will bounce and flow
with the fluid. The critical velocity for the bouncing model was
estimate by Friedlander et al. [24],
0.5
2 1 2
ucr = 2 (1)
m
where is surface potential energy, is the restitution
coefficient, and m is the mass of the particle. The surface
potential energy is defined as:
dH
= (2)
12zo
where H is the Hamaker constant, zo is the distance between the
particle and the wall at steady state, and d is the diameter of the
hydrate particle.
The heat transfer associated with phase change during
hydrate formation is analogous to a convective heat transfer
problem with initial superheat boundary conditions. The
hydrates resemble freezing ice droplets in a flowing stream.
Several analytical methods have been used to solve the
classical phase change problem with initial boundary
conditions [25-29]. This paper expands on integral solutions
presented by Wang and Cheng [30] by taking into account the
growth of the hydrate particle using the bouncing model
proposed by Jassim et al. [19]. The problem is formulated as
phase change of a two-phase fluid using the integral solution of
a boundary layer. The temperature variation is assumed to be
axisymmetric. The heat conduction equation for the hydrate
layer is:
1 T 1 T
r = (3)
r r r t
where r is the internal radius of the pipe, T is the internal
temperature of the fluid, and t is time. The transient term is
reduced using Leibnitzs rule and Eq. (3) is written as:
r
r
rT ( r,t )r + rf To f
t rf t
(4)
T (r,t ) T (rf ,t )
r rf =0
r r
where rf is the hydrate radius and To is the initial temperature of of Stefan number), heat transfer ratio (HTR) and dimensionless
the fluid in the pipe. Non-dimensionalizing Eq. (3) using time are introduced in Eq. (5). These terms are given as:
r
lv
=
T To
Tf To
and
rf

= r dr , Eq. (4) is reduced to: Ph =
k (T f Ta )
(12)

( r f , t ) h2
d
dt
dr
dt
( r , t )
+ rf f r
r
rf
r
=0 (5) hr =
h1
(sup 1) (13)

where is the dimensionless temperature, is the thermal h12t
diffusivity of the fluid, and Tf is the final temperature of the = (14)
k 2 Ph
fluid. Assuming an axisymmetric temperature distribution, the
dimensionless temperature can be represented as a logarithmic where Ph is the phase change number, hr is the HTR, and is
function of r: the dimensionless time.
ln ( r / rf ) Using the non-dimensional parameters (Eq. (12-14)) and
= a+b the boundary conditions (Eq. (7-9)) each term in Eq. 5 can be
ln ( Ro / rf ) (6)
reduced. Integrating,
where a and b are proportionality constants, and Ro is the r2 r 2

external radius of the pipe. = (t ) f

2 2 (15)
Using the following boundary conditions, Eqs. (4) and (6)
can be solved as: The first term in Eq. (5) is written as:

(7) d d drf dr
( rf , t ) = 1 = = rf f (16)
dt drf dt dt
( r,t)) h
= 1 ( r,t ) (8) Similarly, the third term in Eq. (5) is written as:
r k
Bi lv drf
( rr , t ) ) lv dr f Bir (t ) f
= h1 (T f Ta ) dt
r k (T f To ) dt
lv h2 Bi f lv 3 (17)
h f ( sup ) h 2 T T p
h2

k
( )
sup ( rf , t ) +
2 (T f To )
u 3p (9) Bi (t ) + u
( f a )
1 1

where h1 and h2 are the external and internal convective heat Substituting Eqs. (15) and (17) in Eq. (5) and rearranging
transfer coefficients respectively, k is the thermal conductivity,
is the density of the fluid, lv is the latent heat of Bi f Ph drf
2rf +
vaporization, sup is the dimensionless superheat temperature, h1 dt
and up is the particle velocity. The fluid flow stream inside and Bi lv
outside the pipeline are defined by the following Biot numbers. Bir (t ) Bi f hr + f u 3p = 0 (18)
h1 2 (T f Ta )
The internal Biot number will vary based on the thickness of
the hydrate layer, while the external Biot number is assumed to
be constant in this study. This allows the model to be extended The effect of non-dimensional time is introduced in Eq.
to any pipe dimension following similar boundary conditions. (18) by using the following relationship:
drf dBi f d drf
h2rf = (19)
Bi f = dt d dt dBi f
k (10)
Substituting the expression for (Eq. (7)) in Eq. (18) the
hr final expression for the effect of non-dimensional time on Biot
Bir = 1
k (11) number can be written as:

Here Bif and Bir are the internal and external Biot numbers Ph h13 dBi f
rf Bi f 3 +
respectively. h1 k Ph d

The Biot number and dimensionless time are useful to (20)

*
analyze the growth of the hydrate layer for ease of comparison
Bir
Bi U
+ Bi f hr + f = 0
with experimental data and extending the model to various Bir
+
1 h1

hydrate structures. To further simplify the solution for the non- ln ( Bir / Bi f ) ln ( Bir / Bi f )
2

dimensional temperature, the phase change number (reciprocal
where U* =
lv
u 3p . and the internal radius of the pipe. The model assumed that the
2 (T f Ta ) thickness of the pipe has minimal effect on the heat transfer
between the flowing stream and the external fluid.
The growth of the hydrate layer is dependent on kinetic
energy of the particle. Using the bouncing model proposed by
Jassim et al. [19], the particle size of the hydrate can be TABLE I. VARIABLES IN NUMERICAL MODEL
expressed as:
Volumetric density of hydrate 302.4 kJ/mol
n ( r f , t )
t
+
r f
( I ( r f ,t ) n ( r f ,t ) ) = 0 (21) Thermal conductivity R134a 4.35 W/mK
Latent heat of hydrate formation 1996 kJ/kmol
where n is the particle size distribution and I is the growth rate Vapour molar density 0.067 mol/m3
of a single particle which can be expressed as:
Reaction surface area 0.0000375m2
1 dm Fugacity at equilibrium
I ( rf , t ) = (22) temperature 11.9 Pa
4 rf2 h dt
Kinematic constant 3600
The hydrate particle velocity is based on an equation of Sherwood number 2
motion taking into account the free stream velocity of the fluid
and can be written as: Molecular weight 102.3 mol/kg
Molar fraction at equilibrium (yeq) 2.49x10-3
u p = ( u g v ) (1 exp ( t / v ) ) (23)

hdh2Ch The effect of varying the HTR between the fluid in the
where v = is the residence time. In relation to the environment and R134a/water mixture is shown in Fig. 2. The
18
onset region in Fig. 2 is the region where hydrate nucleation
bouncing model, the particle will remain on the surface if the
takes place with no significant growth in the hydrate. An
distance of the particle from the wall is less than the boundary
increase in the HTR reduces the thickness of the hydrate layer
layer thickness determined by the critical velocity in Eq. (1).
as a function of dimensionless time. At the highest HTR (10),
The particle velocity in Eq. (23) is used to determine the
the maximum thickness is approached in a short time as a
kinetic energy of the particle. The diameter of the hydrate is
estimated from the previous time step by estimating the growth result of rapid cooling of the fluid beyond the hydrate forming
rate with a combination of mass transfer and kinetic reaction of temperature. Increasing the HTR reduces the thickness of the
the particle developed by Jassim [31], hydrate layer.
The heat transfer coefficient of the fluid flowing in the pipe
Asmw ( f feq ) nvShDmw y yeq can be increased to reduce the possibility of the hydrate,
rp rp = + (24)
2rp
3000h leading to blockage in the pipe. Although the outer heat
l
i i1
i1
transfer coefficient cannot be controlled in practice, reducing
the heat transfer coefficient of the surrounding fluid can
where f is the fugacity, is the kinematic constant, Sh is the achieve a similar result. The thickness of the hydrate layer is
Sherwood number for mass transfer, nv is the vapor molar dependent on the variation of the phase change number in the
density, mw is the molar mass, y is the molar fraction, and As is transient flow regime (Fig. 3). The phase change number can
the reaction surface. The first term on the right side of Eq. (24) be varied by either the superheat temperature or the HTR. The
is the reaction kinetics component of the growth of the particle, driving force for the hydrate formation is the temperature
while the second term is associated with mass transfer. difference between the fluid and phase transition temperature.
The system of Eqs. (1), (20), (23), and (24) are solved using
a Newton-Rapson algorithm to determine the growth of the
hydrate in terms of the dimensionless Biot number at each time
step. The numerical model was used to investigate how key
dimensionless quantities and physically measureable quantities
such as temperature (fluid, environment, and interface
temperature), and pipe radius affect the hydrate growth.

III. RESULTS AND DISCUSSION

The predicted result is compared against experimental
conditions reported by Dorstewitz and Mewez [15] for R134a
hydrates. The assumptions for both the numerical and
experimental studies are shown in Table I. The parameters
which affect the hydrate formation include the temperature of
the surrounding environment, fluid temperature, heat transfer
Fig. 2. Effect of HTR on hydrate growth (external temperature -3oC)
characteristics of the fluids internal and external to the pipeline,
As the superheat temperature increases, the heat transfer to the
hydrate layer also increases, reducing the growth rate of the
hydrate layer. The hydrate layer thickness increases by a factor
of 4 when the phase change number is reduced by 25%, as the
hydrate layer approaches steady state. The phase change
number has minimal effect on hydrate formation when the
hydrate thickness approaches steady state.
Fig. 3. Effect of phase change number on hydrate formation (pipe diameter
0.3m and external temperature -3oC)
The effect of changing environmental conditions for
specific internal flow is shown in Fig. 4. Changing the external
Biot number can be accomplished by a variation in the heat
flux or thermal conductivity. The peak of each curve is the
instance at which the hydrate completely fills the pipeline
diameter and impedes the flow of the fluid. At very low Biot
numbers, it takes a longer time for the hydrate to completely
impede the fluid flow. This will indicate that pipes with
external surfaces with very high thermal conductivity could
reduce the hydrate formation in the pipe. The regions beyond
the peak in each curve in Fig. 4 are non-physical, where the
model is no longer applicable, since the hydrate cannot grow
beyond the actual pipe diameter.
The numerical model was validated by comparing the
results with experimental results reported by Dorstewitz and
Mewez [15] for R134a hydrates (Fig. 5). The comparison
shows that the predicted model in this paper is more accurate at
higher dimensionless times.
Fig. 4. Effect of change in external Biot number on hydrate formation (fixed
inlet temperate 5oC and pipe diameter of 0.15m)
The initial discrepancy at earlier times occurs because the
model does not include the initial velocity of the hydrate
particle in the results of Dorstewitz and Mewez [15].
Performing a sensitivity analysis of the result, about an 8%
error is observed in comparison to the experimental results at
early dimensionless time. The difference between numerical
and experimental results is about 1% as the dimensionless time
approaches 0.5. Accounting for the hydrate particle velocity in
the model is important in predicting the hydrate formation as in
Fig. 5. The rate at which the hydrate blocks the flow stream is
better predicted with the present model.
Fig. 5. Comparison of predictions with experimental data for R134a hydrates
The effect of fluid flow temperature on the hydrate
formation is shown in Fig. 6. In order to compare the results
with experimental data in this case, similar operating
conditions used by Dorstewitz and Mewez [15] were adopted
in the analysis. The temperature of the internal fluid is fixed at
5oC and the external temperature is varied between -3 and -
20oC in Fig. 6. As expected, the hydrate formation occurs faster
in the colder environment. The rate of hydrate formation
between temperatures of -3 and 10oC increased by about 53%.
The increase in the rate of hydrate formation between -10 and -
20oC is less significant with only a 7.3% variation. For the
same internal fluid, hydrate mitigation techniques at extreme
temperatures will be more expensive with limited design
modifications to reduce the cost of mitigation techniques.
Fig. 6. Effect of changing external environment temperature on hydrate
formation (supply temperature of 5oC and a pipe diameter of 0.15m)
A change in the superheat temperature is a function of the
interface temperature between the hydrate and fluid boundary.
The internal and external fluid temperatures are kept constant,
while the interface temperature is varied in Fig. 7. The
maximum variation in Biot number is about 0.5%, which
indicates that interface temperature has no significant effect on
the rate of hydrate formation. The dimensionless superheat
Tsup Ta
temperature sup = is used to determine the superheat
Tf Ta
temperature based on the HTR. The HTR in Eq. (13) is used to
determine the dimensionless superheat temperature. The
relationship between the dimensionless superheat temperature
and the HTR clearly indicates that the dimensionless
temperature is dependent on the ratio of the heat transfer
coefficient between the internal and external fluids.
Fig. 7. Effect of superheat temperature on hydrate formation (external
temperature of -3oC and pipe diameter of 0.15m)
The effect of the pipe radius on the hydrate formation
is shown in Fig. 8. A change in pipe radius between 15 and 60
mm increased the rate of hydrate formation by about 85% for
the range of dimensionless time in this study. As the radius of
the pipe increases, there is an increased flow of heat to the
surrounding fluid due to the increased surface area in contact
with the environment.
Fig. 8. Effect of changing pipe diameter on hydrate formation (external
temperature of -3oC)
The effect of internal radius on the rate of hydrate
formation increased at larger diameters, as a 52% increase is
observed from a change of 50 and 60mm. This indicates an
optimal diameter can be determined using this model for a
specific region of operation.
IV. CONCLUSIONS
A new predictive model in this paper can be used to
account for hydrate formation in pipelines. The semi-analytical
models use reduced computational resources compared to a
CFD analysis. The numerical model can be extended to handle
mixtures of hydrates by taking into account mixture densities.
If the temperature of the environment reduces below -10oC,
modifying any of the thermodynamic factors produces less
impact on reducing the hydrate formation. Minimizing the
radius of the pipe reduces the rate of hydrate formation in the
pipe. The numerical model presented in this paper provides a
useful predictive tool to analyze hydrate formation in real time,
as this will allow for integration into existing hydrate
monitoring devices.
ACKNOWLEDGEMENTS
Financial support from the Natural Sciences and
Engineering Research Council of Canada (NSERC), and
Research and Development Cooperation of Newfoundland and
Labrador, is gratefully acknowledged.
REFERENCES
[1] J. Parent, P. Bishnoi, Investigations into the nucleation behaviour of
methane gas hydrates, Chem. Eng. Commun. 144 (1996) 51-64.
[2] J.D. Lee, R. Susilo, P. Englezos, Kinetics of structure H gas hydrate,
Energy Fuels. 19 (2005) 1008-1015.
[3] H. Ganji, M. Manteghian, K. Sadaghiani Zadeh, A kinetic study on
tetrahydrofuran hydrate crystallization, J. Chem. Eng. Japan. 39 (2006)
401-408.
[4] D. Kashchiev, A. Firoozabadi, Nucleation of gas hydrates, J. Cryst.
Growth. 243 (2002) 476-489.
[5] P. Englezos, N. Kalogerakis, P. Dholabhai, P. Bishnoi, Kinetics of
formation of methane and ethane gas hydrates, Chem. Eng. Sci. 42
(1987) 2647-2658.
[6] P. Skovborg, P. Rasmussen, A mass transport limited model for the
growth of methane and ethane gas hydrates, Chem. Eng. Sci. 49 (1994)
1131-1143.
[7] J. Herri, J. Pic, F. Gruy, M. Cournil, Methane hydrate crystallization
mechanism from insitu particle sizing, AICHE J. 45 (1999) 590-602.
[8] N. Gnanendran, R. Amin, Modelling hydrate formation kinetics of a
hydrate promoterwaternatural gas system in a semi-batch spray
reactor, Chem. Eng. Sci. 59 (2004) 3849-3863.
[9] O O. Kwon, S. Ryou, W. Sung, Numerical modeling study for the
analysis of transient flow characteristics of gas, oil, water, and hydrate
flow through a pipeline, Korean J. Chem. Eng. 18 (2001) 88-93.
[10] R.H. Nemoto, J.L. Balio, Modeling and simulation of severe slugging
with mass transfer effects, Int. J. Multiphase Flow. 40 (2012) 144-157.
[11] J. Niessner, S.M. Hassanizadeh, Non-equilibrium interphase heat and
mass transfer during two-phase flow in porous media-theoretical
considerations and modeling, Adv. Water Resour. 32 (2009) 1756-1766.
[12] V.S. Shagapov, R. Urazov, N. Musakaev, Dynamics of formation and
dissociation of gas hydrates in pipelines at the various modes of gas
transportation, Heat Mass Transfer. 48 (2012) 1589-1600.
[13] E. Tolubinskiy, An integral method of solution of the general heat and
mass transfer problem, Int. J. Heat Mass Transfer. 9 (1966) 1471-1488.
[14] J. Mennig, M. ziik, Coupled integral equation approach for solving
melting or solidification, Int. J. Heat Mass Transfer. 28 (1985) 1481-
1485.
[15] F. Dorstewitz, D. Mewes, The influence of heat transfer on the
formation of hydrate layers in pipes, Int. J. Heat Mass Transfer. 37
(1994) 2131-2137.
[16] M.J. Vaze, J. Banerjee, A modified heat transfer correlation for two-
phase flow, Heat mass transfer. 47 (2011) 1159-1170.
[17] K. Lin, P. Wei, S. Hsiao, Unsteady heat conduction involving phase
changes for an irregular bubble/particle entrapped in a solid during
freezing-An extension of the heat-balance integral method, Int. J. Heat
Mass Transfer. 52 (2009) 996-1004.
[18] T. Elperin, A. Fominykh, Heat transfer during gas hydrate formation in
gas-liquid slug flow, Heat Mass Transfer. 31 (1996) 179-183.
[19] E. Jassim, M.A. Abdi, Y. Muzychka, A new approach to investigate
hydrate deposition in gas-dominated flowlines, J. Natural Gas Sci. Eng.
2 (2010) 163-177.
[20] C.P. Ribeiro Jr, P.L. Lage, Modelling of hydrate formation kinetics:
State-of-the-art and future directions, Chem. Eng. Sci. 63 (2008) 2007-
2034.
[21] G. Naterer, Coupled liquid film and solidified layer growth with
impinging supercooled droplets and joule heating, Int. J. Heat Fluid
Flow. 24 (2003) 223-235.
[22] A. Aiyejina, D.P. Chakrabarti, A. Pilgrim, M. Sastry, Wax formation in
oil pipelines: a critical review, Int. J. Multiphase Flow. 37 (2011) 671-
694.
[23] B. Tohidi, A. Chapoy, J. Yang, Developing a hydrate-monitoring
system, SPE Projects Facilities & Construction. 4 (2009) 1-6.
[24] S.K. Friedlander, Smoke, Dust, and Haze, Oxford University Press New
York, 2000.
[25] H.S. Carslaw, J.C. Jaeger, Conduction of heat in solids, Oxford:
Clarendon Press, 2nd Edition, 1959.
[26] T.R. Goodman, The heat-balance integral and its application to problems
involving a change of phase, Trans. ASME. 80 (1958) 335-342.
[27] F. Kreith, F. Romie, A study of the thermal diffusion equation with
boundary conditions corresponding to solidification or melting of
materials initially at the fusion temperature, Proc. Phys. Soc.Ser B. 68
(1955) 277-291.
[28] V.J. Lunardini, Heat Transfer in Cold Climates, Van Nostrand Reinhold
Company New York, 1981.
[29] L. Tao, Free-boundary problems with radiation boundary-conditions,
Appl. Math. 37 (1979) 1-10.
[30] C. Wang, P. Cheng, A multiphase mixture model for multiphase,
multicomponent transport in capillary porous media-I. Model
development, Int. J. Heat Mass Transfer. 39 (1996) 3607-3618.
[31] E.I. Jassim, Locating Hydrate Deposition in Multiphase Compressed
Natural Gas Flow Lines Using Computational Fluid Dynamics
Techniques. PhD Thesis, Memorial University of Newfoundland, St.
John`s (2008).