Professional Documents
Culture Documents
H I GH L IG H T S
A R T I C LE I N FO A B S T R A C T
Keywords: Supersonic separation is a novel technology. A multi-fluid slip model for swirling flow with homogenous/het-
Supersonic separator erogenous condensation and evaporation processes in the supersonic separator was built to estimate the se-
Condensation and evaporation paration efficiency. This model solves the governing equations of compressible turbulent gas phase and dispersed
Multi-fluid model homogenous/heterogenous liquid phase considering droplet coalescence and interphase force. Its prediction
Natural gas dehydration
accuracy for condensation and swirling flows was validated. Then, the flow field, slip velocity and droplet
Separation efficiency
trajectory inside the separators with different swirl strengths were investigated. The maximum values of radial
slip velocity are 29.2 and 8.26 m/s for inlet foreign droplet radius of 1.0 and 0.4 µm. It means the larger foreign
droplet has a better condensation rate. However, the residence time of larger foreign droplet in core flow is
shorten. Thus, the inlet radius of foreign droplet has to be moderate for best separation efficiency. Finally, the
dehydration performances of separator were evaluated. The optimal radius of inlet foreign droplet to maximize
the dehumidification and efficiency was found. For the separator with swirl strength of 22%, the optimal radius
is 0.85 µm at inlet pressure of 250 kPa, where the maximum dew point depression is 42.41 °C and the water
removal rate is 87.82%.
⁎
Corresponding author.
E-mail address: wangchao@tju.edu.cn (C. Wang).
https://doi.org/10.1016/j.applthermaleng.2020.115074
Received 12 August 2019; Received in revised form 9 February 2020; Accepted 11 February 2020
Available online 11 February 2020
1359-4311/ © 2020 Elsevier Ltd. All rights reserved.
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
Nomenclature Greek
system because of the low temperature and short residence time inside water dew point. Wen et al. [13] compared the temperature and pres-
the supersonic separator [10]. Although many works have been done sure profiles obtained by FLUENT with phase envelope diagrams by
for optimizing the structure and performance of supersonic separator, HYSYS and found the supersonic separators is suitable for liquefied
the actual dehydration performance is still far from being satisfactory, natural gas. Hammer et al. [14] investigated CO2 removal of supersonic
as the flow inside supersonic separator is quite complex which includes separator from dry 2.98% CO2 exhausts by using Aspen HYSYS. Ni-
condensation, evaporation, vorticity, turbulence, shockwave and in- knam et al. [15] used Aspen HYSYS to predict the 3S unit performance
terphase slip. by conventional unit operations, then analyzed the performance of
The goal of supersonic separator is to remove condensable vapors nozzle and proposed an efficiency model for supersonic separation by
from the gas flow and achieve a lower dew point or strip the gas of integrating COMSOL, HYSYS and MATLAB software [16]. Arinelli et al.
heavy hydrocarbons. The dew point depression and water removal rate [17] developed two steady state Unit Operation Extensions (UOE) and
of the supersonic separator are really desired. Currently, there are two the vapor-liquid equilibrium (VLE) envelopes for HYSYS 8.8 to simulate
approaches for predicting its separation performance, the computa- supersonic separator and membrane permeation for treating humid
tional fluid dynamics using Fluent software [4] and the process simu- natural gas with 44%mol CO2. Bian et al. [18] proposed a novel gas
lation conducted in Aspen HYSYS [11]. liquefaction process and the liquefaction efficiency under different
For process simulation, Karimi and Abdi [12] investigated the flow condition were calculated by using MATLAB and HYSYS. Rezakazemi
field of single-phase gas in a Laval nozzle without swirl by using Aspen et al. [19] carry out a series of process simulation by Aspen HYSYS to
HYSYS, and found the gas could potentially be dehydrated to very low compare recovery of ethane from typical JT Valve, Turbo-Expander and
2
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
Twister Technology. Brigagão et al. [20] developed a new CCS unit by laboratory tests, and leads to the heterogeneous condensation. This
using a supersonic separator based on HYSYS 8.8. CO2 emissions can be process can effectively reduce the free energy barrier and promotes the
reduced by 43% with economic performance. de Medeiros et al. [21,22] condensation appearance. Ma et al. [40] experimentally investigated
proposed UOE module for estimating sound speed of two-phase systems the effect of foreign droplets on vapor condensation by using droplet
in supersonic reactors for multiphase multi-reactive equilibrium system enlargement according to heterogeneous condensation theory. Ding
via HYSYS. However, it is recognized that their results were not vali- et al. [41] carried out a two-dimensional multi-phase turbulence model
dated against experimental data. for homogeneous/heterogeneous condensation to investigate the effect
For CFD simulation, Malyshkina [23] numerically investigated the of condensation on mass flow rate of Laval nozzle without swirl. Ni-
single-phase swirling flow inside a supersonic separator and analyzed knam et al. [42] experimentally obtained dehydration of low-pressure
effects of Mach number and temperature on recovery of pressure. Wen gas in supersonic separator with heterogeneous condensation of water
et al. [24,25] studied the flow characteristics and mass flux inside a droplets containing small foreign droplet and built a two-phase Euler-
supersonic separator with and without swirls. Wang and Hu [26] in- Euler model in 2D asymmetric domain for analyzing the dehydration
vestigate the effects of swirling generator and drainage structure where efficiency. However, they did not consider homogenous nucleation
the RNG k-ε turbulence model was utilized. Liu [27] established a process and the drag model was not provided. Shooshtari and Shahsa-
three-dimensional single-phase numerical model considering the com- vand [43] built a heterogeneous condensation model without inter-
pressible and strong swirling effect to study the flow field and to opti- phase slip for optimizing geometry of the Laval nozzle which was va-
mize separation performance. Alnoush and Castier [28] proposed a lidated by Dykas’s predictions [44]. Bian et al. [45] built a physical
shortcut one-dimensional single-phase numerical modeling for the model for homogenous and heterogeneous condensations in a Laval
conceptual preliminary design of supersonic separator. nozzle. However, the interphase slip velocity was ignored which is not
On this basis, two-phase models based on the discrete particle suitable for the swirling flow in a supersonic separator.
method (DPM) were built. Wen et al. [29] and Yang et al. [30] studied As mentioned above, there is a lack of CFD model for accurate
natural gas flow characteristics and particle separation behavior by prediction of liquid droplets trajectories inside supersonic separator
using DPM model with particle diameter constant of 1.4 µm. Liu et al. with combination of swirling, nucleation, growth, evaporation, slip,
[31] also predicted the particle behavior inside a supersonic separator collision and coalescence in respect of the operation optimization of the
employing the DPM model, where the particle diameter range is from supersonic separator. This paper proposed a novel multi-fluid (slip)
10 to 50 µm, However, the typical particle size in the supersonic se- model for the swirling flow with the homogenous/heterogenous con-
parator is 0.1–2 µm [32] from Twister BV. Jiang [33] studied the gas densation and evaporation processes in supersonic separator. The
flow field and CO2 droplet trajectory inside a supersonic separator mathematical modeling and performance assessment of present multi-
employing DPM model. The size of CO2 droplet is set to be a constant of fluid model have been conducted. It provides an effective method to
1.25 µm. However, these researches do not consider the complex nu- analyze the effects of homogeneous/heterogeneous condensation with
cleation-condensation process of condensable vapor and the evapora- enhanced heat transfer coefficient due to strong swirl on the dehydra-
tion process of droplet inside the supersonic separator. tion efficiency of supersonic separator.
Take into account the condensation process in the supersonic flow,
Ma et al. [34] built an Eulerian two-fluid model to investigate homo- 2. Supersonic separator
genous condensing flow in a Laval nozzle. Shooshtari and Shahsavand
[35] provided a new theoretical approach based on mass transfer rate to In this paper, the supersonic separator is composed of a central body
predict liquid droplet growth of homogenous condensation of binary and some vanes were designed, as shown in Fig. 1. The annular channel
mixtures in a Laval nozzle. Ding et al. [36] proposed an analytical so- between the central inner body and the shell wall forms a supersonic
lution to predict Wilson point of homogeneous nucleation inside the nozzle section, a cyclonic separation section and a diffuser with the dry
nozzle. Castier [37] conducted several numerical simulations for the and wet gas outlets [24]. The swirling flow is generated by several
natural gas partial condensation inside a Laval nozzle without swirl. static vanes with specific turning angle at the subsonic section which
Dykas [38,39] built an in-house CFD code with 3-D RANS equations for can control the flow swirl strength.
single-fluid no-slip model to predict the wet steam and moist air con- Several parameters related to the separation efficiency and hydra-
densation process within a Laval nozzle. Wen et al. [4] developed a tion performance were defined. The pressure loss ratio γ is expressed as
numerical model of homogenous nucleation to predict CO2 condensa-
tion because of condensation inside supersonic flows. The condensation pin − pout
γ=
model predicted 18.6% of condensed liquid fraction. pin (1)
Actually, raw natural gas usually contains foreign droplets before
where p is the pressure of the fluid. The superscripts in and out re-
entering into the supersonic separator in practical operation and
present inlet and outlet conditions.
3
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
The recovery of dry gas at the outlet is defined as where the total energy is E and hlg is the latent heat, J kg−1. The ef-
cpg μ
fective turbulent thermal conductivity is λ eff = λ g + Pr tu . The devia-
Q dry Q in − Qwet tu
ψ= × 100\% = × 100\% toric stress tensor (τij )eff is defined by
Q in Q in (2)
where Q in , Q dry and Qwet represent the mass flow rates at inlet, dry and ∂u¯ gj ∂u¯ gi ⎞ 2 ∂u¯ gk
(τij )eff = μ ⎛⎜ + ⎟ − μeff δij
wet gas outlets. ⎝ ∂x i ∂x j ⎠ 3 ∂xk (8)
The dew point depression ΔTd is defined as [40],
where the effective turbulent viscosity is μeff = μ + μtu and the turbu-
ΔTd = Tdin − Tdout (3) lent viscosity μtu is described by
divided into two parts, ΔTd1 due to expansion (without phase change) sipation rate, s−1, respectively. The strain rate magnitude is
and ΔTd2 due to separation (with phase change), namely Ω = 2Si,j Si,j and the component of mean strain rate Si,j is computed by
ΔTd = ΔTd1 + ΔTd2. The effective dew point depression is ΔTd2 due to
the condensation and separation. Si,j =
1
2 ( ∂u¯j
∂x i
+
∂u¯i
∂x j ). The coefficient a* is a low-Reynolds number cor-
The water removal rate ηv at dry gas outlet is calculated by rection. a1 is a constant 0.31. F2 is the blending function.
The species transport equation for gas phase is expressed by
χvin − χvdry
ηv = × 100\%
χvin (4) ∂ ∂ ∂ ⎛ ∂cs ⎞
(αg ρg cs) + (αg ρg u j cs) = ⎜Ds α g ρg ⎟ + Scs
∂t ∂x j ∂x j ⎝ ∂x j ⎠ (10)
where, χ is the mole fraction of water vapor and the subscript v re-
presents the water vapor. where cs is the species mass fraction and Ds is mass diffusion coefficient.
Scs is source term in species transport equation. If the species is water
3. Mathematical model vapor, Scs = −(mhom + mhet).
The turbulence kinetic energy and specific dissipation rate for gas
3.1. Multi-fluid model for condensation-evaporation swirling flow phase are expressed as:
4
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
where FDi,hom and FDi,het are the components of drag force between the where TR = Tg/Tc is the reduced temperature and Tc is critical tem-
continuous and the dispersed phases. perature (647.3 K). The dew point temperature Td is deduced by [48]
⎝ 1 − 0.5767 ⎠ (30)
respectively. The supersaturation of vapor is Ss = pv/ps(Tg), where ps
(Tg) is the water vapor saturation pressure at temperature Tg. where the latent heat of water hlg at the normal boiling point hlg0 is
The critical nucleation radius rc of the droplet is described by 2262.2 kJ·kg−1.
2σ
rc = 3.3. Droplet coalescence and drag model
ρl Rv Tv ln(Ss ) (20)
where, Rv represent specific gas constant of water vapor. The mean The mean collision frequency f of one heterogeneous foreign droplet
radii rhom and rhet are expressed as (collector) with surrounding smaller homogeneous droplets is described
1 1
by (O’Rourke [50])
3α 3 3α 3
r hom = ⎛ hom ⎞ , r het = ⎛ het ⎞
⎜ ⎟ ⎜ ⎟ f = nhom π (r het + r hom )2urel (31)
⎝ 4πn hom ⎠ ⎝ 4πnhet ⎠ (21)
where urel = |uj,het − uj,hom| is relative velocity between homogeneous
Thus, the liquid mass changing rates mhom and mhet are calculated
and heterogeneous droplets. After the collisions, these smaller homo-
by
geneous droplets are collected by heterogeneous droplet, and the het-
4πrc3 2 dr hom 2 dr het
erogeneous droplet radius increases up gradually. It is necessary to
mhom = Jρl + nhom ρl 4πrhom , mhet = nhet ρl 4πrhet
3 dt dt (22) consider the effect of drag force in two-phase momentum equations for
the swirling flow with strong centrifugal force. For smaller Reynolds
The growth rate of droplet (both condensation and evaporation) is
numbers, the molecular viscous force dominates [51]. Thus, according
expressed as
to the Stokes’ law, the drag force FD is given by [52]
3
∑ ak ⎯→
⎯ 9μg ⎯→
⎯ 9μg
dr FD,hom = αhom (→
ug − →
uhom ), FD,het = αhet (→
ug − →
uhet )
= k= 1 (Tr − Tv ) 2
2rhom Cc 2
2rhet Cc
dt ρl hlg (23)
(32)
where ak is heat transfer coefficient. The subscript k represents different where Cc is the empirical Cunningham slip correction factor which is
species in gas phase. The droplet surface temperature Tr is calculated by calculated by
rc
Tr = Td (pv ) − D T Cc = 1 + 2Kng (1.257 + 0.4e−1.1 (2Kng) )
(33)
r (24)
where DT is subcooling, =Td(pv) − Tg. Td(pv) is the dew point tem-
perature at local water vapor pressure. 4. Experimental validation
The flow inside the supersonic separator is strongly rotational and
turbulent. The centrifugal acceleration can achieve 106–107 m s−2 [24]. 4.1. Numerical implementation
The velocity slip between the gas and liquid phase has to be taken into
account, for purpose of calculating the heat transfer coefficient accu- All the cases were simulated by ANSYS FLUENT, while the gov-
rately. The heat transfer coefficients ak between the water droplet and erning equations and the source terms were performed with C code by
each gas species are calculated by the User-Defined-Scalar (UDS) and User-Defined-Function (UDF) in-
terfaces. The density-based solver with implicit formulation and the
λk 1
ak = (2 + 0.6Re0.5
p Prg
0.33)
Roe-FDS for flux type were utilized for calculating governing equations
2r 1 + 2 8π Γk
1.5 Pr k
× 1 + Γk
Knk (25) of multi-fluid flows including the gas and liquid phases. The second-
order upwind scheme was employed for discretization. The gradients of
where Rep, Г, λ and Kn are Reynolds number, specific heat ratio,
flow variables for spatial discretization were calculated by the least
thermal conductivity and Knudsen number, respectively. For the water
squares cell based gradient evaluation. For turbulence model, the SST k-
vapor, Kn is expressed as
ω model was selected as the viscous model of gas phase. The structured
3μ v Rv Tg quad-map meshes were utilized for better convergence. For inlet
Knv = boundary conditions, the total temperature and pressure of fluid were
4pv r (26)
given. The static pressure was set at the outlet. The concentration nhetin
The relative Reynolds number Rep, het is described as of inlet foreign droplet with a given radius rhetin were specified at the
inlet. The constant-temperature condition which is equal to inlet gas
Re p, het = 2ρg r het |→
uhet − →
ug| μg = 2ρg r het |Vs| μg (27) temperature was enforced at walls. The solutions of density ρg, velocity
where Vs is the slip velocity, =uhet-ug. Besides, the liquid surface ten- ugj, energy E, turbulence kinetic energy k, specific dissipation rate ω of
sion σ is calculated by [47] the gas phase, the droplet concentration nhom and nhet, the liquid phase
effective density ρhom and ρhet, the liquid velocity uj,hom and uj,het were
σ = (85.27 + 75.612TR − 256.889TR2 + 95.928TR3 ) × 10−3 (28) converged to normalized residual mean square of 10−5.
5
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
Fig. 2. The experimental apparatus for the nozzle and supersonic separator.
6
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
(b) The profiles of pressure ratio at the nozzle center downstream of nozzle throat
Fig. 3. Comparisons of CFD results with experimental data reported by Wyslouzil.
7
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
8
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
⎨ Γp(1 + Γ) Γ Kl ⎝ Kl ⎠ ⎦ ⎬ previously mentioned, both dry gas outlet and wet gas outlet are
⎩ ⎣ ⎭ (35)
100 kPa for Case 1 and Case 3. The heterogenous condensation is
where β is gas-liquid mass ratio. Gg,ref = Tref Rg/ρrefΓ−1.
ρl,ref and pref are dominant and the homogenous condensation can be ignored in these
chosen as 1000 kg/m3 and 1 bar. two cases.
It is found that the results of Eqs. (34) and (35) are almost same. The contour of the liquid mass changing rate mhet for Case 1 when
9
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
mass changing rate mhet is from −936.7 to 958.3 kg m−3 s−1 (max-
imum latent heat release rate is 2.51 × 109 J m−3 s−1). The liquid mass
changing rate at the near wall cells are shown in Fig. 13(b). It is found
that the evaporation process occurs in two parts, the near-wall heating
zone and pressure recovery zone.
For the purpose of demonstrating the condensation and evaporation
processes, the distributions of the droplet radius and the vapor mass
fraction along the axial direction are plotted in Figs. 14 and 15. The
radial positions xr of Line 1-Line 3 are 7 mm, 8 mm and 9 mm, re-
spectively. Fig. 14 shows radius distributions of heterogenous
droplet along Line 2 at different total pressures when inlet foreign
droplets rhetin = 1.0 μm, and ρhetin = 0.01 kg m−3.
It indicates that the heterogenous droplet size starts growing rapidly
at the throat. Along with the increasing of heterogenous droplet size, its
changing rate becomes smaller. The heterogenous droplet size will
reach a maximum at the location of shock wave, and then begin to
decrease due to the evaporation. The heterogenous droplet size grows
to 1.44, 1.83 and 2.26 μm at xa = 91.35, 174.50 and 227.40 mm for
Fig. 10. The swirl strength effect on mass flow rate in the supersonic separator.
p*in = 167, 200 and 250 kPa.
The mass fraction of the water vapor in the gas phase also can re-
p*in = 167 kPa is plotted as Fig. 13(a). It reveals that both condensation flect the changes of condensation and evaporation. Fig. 15 shows the
and evaporation processes occur inside the separator. The global range results of mass fraction of water vapor for Case 3 with rhetin = 1.0 μm
of the liquid mass changing rate mhet is from −402.3 to 712.1 kg m−3 and Case 4 with rhetin = 0.4 μm. It is found that the mass fraction of
s−1. According to Fig. 11(a), the shock occurs at axial position water vapor at entrance of drainage will increase with the decrease of
xa = 91 mm which coincides with the evaporation onset. This is be- the inlet foreign droplet size, which means the smaller inlet foreign
cause the fluid velocity begins to slow down and the static temperature droplet radius rhetin is, the worse dehydration performance is. The dif-
recovers to high value downstream of shock wave, and then droplets ferences of performance are attributed to the interphase velocity slip
evaporate into the gas phase gradually. and the uneven distribution of liquid phase.
When the inlet pressure increases to 250 kPa, the rates of both
condensation and evaporation are bigger. The global range of liquid
10
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
(a) at the core region for Case 1: p*in = 167 kPa ( = 0.4)
(b) at the near wall cells for Case 3: p*in = 250 kPa ( = 0.6)
Fig. 13. The contours of the liquid mass changing rate mhet in the condensation and evaporation zones.
5.1.3. Slip velocity and liquid phase distribution of the drag force. Besides, the axial slip velocity along the radial di-
The slip velocity and drag force play a critical role in controlling the rection is from zero (central inner body) to a negative minimum to zero
trajectory of droplet which can determine the separation performance again and then to positive maximum at the shell wall. Comparing
in the cyclone section. As the main factor, the effect of inlet foreign Fig. 16(a) and (c), it indicates that the bigger inlet droplet diameter is,
droplet radius on the slip velocity distribution is discussed firstly. the bigger droplet inertia is, and the larger slip speed will be. Fig. 16(b)
The axial and radial components of slip velocity at different loca- and (d) show the radial component of slip velocity at different loca-
tions of Case 3 and Case 4 are shown in Fig. 16. As shown in Fig. 16(a), tions. As we know, the radial slip speed causes the droplets to separate
the magnitude of axial component of slip velocity decreases gradually from primary gas fluid. It is obvious that the axial velocity of centrifugal
along the flow direction where the maximum velocity lag of the droplet droplet is larger than local velocity of gas phase. The maximum radial
for xa = 90, 140, and 190 mm is 48.2, 33.9 and 24.3 m s−1 as the result slip speed is 29.2 m s−1 for Case 3 and 8.26 m s−1 for Case 4.
11
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
12
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
(a) Axial component of slip velocity for Case 3 (b) Radial component of slip velocity for Case 3
(c) Axial component of slip velocity for Case 4 (d) Radial component of slip velocity for Case 4
Fig. 16. The axial and radial components of the slip velocity at different locations.
Figs. 19 and 20 also illustrate the supersonic separator will provide helpful for designing the separator structure and optimizing the se-
better dehydration performance if the pressure loss ratio increases. The paration performance.
dew point depression ΔTd versus the pressure loss ratio γ is plotted in
Fig. 21. The original data of Fig. 21 is also listed in Table 4. It reveals 6. Conclusion
that the dew point depression ΔTddry increases from 6.91 °C
(ηv = 16.23%) to 52.04 °C (ηv = 93.42%) when the pressure loss ratio A novel multi-fluid model with the droplet coalescence and the in-
increases up to 0.7. The larger the pressure loss ratio inputs, the higher terphase force for the homogenous/heterogenous condensation and
the separation efficiency will be. However, high pressure loss ratio evaporation processes in supersonic separator was built to evaluate the
means high energy consumption. dehydration performance. The flow field, slip velocity profile, droplet
Besides, the effect of swirl strength on the separation efficiency was trajectory and liquid phase distribution with different pressure loss
also investigated. The dehydration performance of supersonic separator ratios, the radii and concentrations of inlet foreign droplets were ana-
B with 45% swirl strength is shown in Table 5. The optimal size of inlet lyzed. The dehydration performances were also investigated further.
foreign droplets for heterogenous condensation is 0.28 µm. Compared The conclusions are as follows:
with the results of separator A with 22% swirl strength in Table 2, it is
found that the optimal radius of inlet foreign droplets will be smaller (1) The complex lambda shock-system with maximum Mach number of
when the separator has a stronger swirl strength. These results are 1.765 and subcooling of 48.76 °C was observed in the cyclone
13
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
(a) Single-fluid (no-slip) model (8% heterogenous +90% homogeneous condensations at xa = 240 mm)
(b) Multi-fluid (slip) model (91% heterogenous + 0% homogeneous condensation at xa = 240 mm)
Fig. 17. Comparisons of streamline and effective density of heterogeneous droplets between single-fluid model and multi-fluid model for Case 3.
Fig. 18. Liquid phase effective density of heterogeneous condensation for Case 4.
Table 2 Table 3
Dehydration performance of supersonic separator A with different rhetin and Dehydration performance of supersonic separator A with different rhetin and
ρhetin of inlet foreign droplets when p*in = 250 kPa (γ = 0.6). ρhetin of inlet foreign droplets when p*in = 167 kPa (γ = 0.4).
rhetin ρhetin nhetin [m−3] pvdry [Pa] Tddry [°C] ΔTddry [°C] ηv rhetin ρhetin nhetin [m−3] pvdry [Pa] Tddry [°C] ΔTddry [°C] ηv [%]
[μm] [kg [μm] [kg
m−3] m−3]
1.0 0.04 9.57 × 1012 86.65 −21.71 51.71 94.90% 1.0 0.04 9.57 × 1012 2059.43 17.80 12.19 19.05%
0.02 4.78 × 1012 130.62 −17.40 47.40 92.31% 0.02 4.78 × 1012 1571.55 13.58 16.42 38.20%
0.01 2.39 × 1012 225.68 −11.48 41.48 86.71% 0.01 2.39 × 1012 1446.73 12.32 17.68 43.11%
0.005 1.20 × 1012 348.99 −6.59 36.59 79.45% 0.005 1.20 × 1012 1483.44 12.69 17.31 41.66%
0.002 4.78 × 1011 736.66 2.63 27.37 56.63% 0.002 4.78 × 1011 1694.78 14.74 15.26 33.35%
0.001 2.39 × 1011 1197.13 9.49 20.51 29.53% 0.001 2.39 × 1011 2113.84 18.22 11.78 16.87%
0.0005 1.20 × 1011 1532.67 13.19 16.81 9.77% 0.0005 1.20 × 1011 2387.42 20.19 9.80 6.13%
2.0 0.01 2.99 × 1011 1319.60 10.93 19.07 22.32% 2.0 0.01 2.99 × 1011 2027.67 17.56 12.44 20.26%
0.5 1.91 × 1013 331.90 −7.16 37.16 80.46% 1.5 7.09 × 1011 1539.66 13.26 16.74 39.45%
0.4 3.74 × 1013 726.40 2.44 27.56 57.24% 0.75 5.67 × 1012 1831.56 15.95 14.05 27.97%
0.3 8.86 × 1013 1418.75 12.02 17.98 16.48% 0.5 1.91 × 1013 2241.32 19.17 10.83 11.86%
0.25 1.53 × 1014 1750.03 15.24 14.76 −3.02% 0.3 8.86 × 1013 2984.91 23.89 6.10 −17.43%
0.2 2.99 × 1014 1999.98 17.34 12.66 −17.74% 0.2 2.99 × 1014 3368.58 25.95 4.05 −32.47%
section, where the maximum tangential velocity was 136.1 m s−1 958.3 kg m−3 s−1 with latent heat release rate of 2.51 × 109 J m−3
and the corresponding centrifugal acceleration was 3.5 × 106 m s−1. When rhetin = 1.0 μm, and ρhetin = 0.01 kg m−3, the hetero-
s−2. genous droplet size grows to 1.44, 1.83 and 2.26 µm for p*in = 167,
(2) The condensation begins to occur at the throat, while the eva- 200 and 250 kPa.
poration appears at the near-wall heating and pressure recovery (3) The magnitude of axial component of slip velocity decreases gra-
zones. The liquid mass changing rate mhet ranges from −936.7 to dually along the flow direction where maximum velocity lag of the
14
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
Table 4
Dehydration performance of supersonic separator A with different pressure loss
ratio γ when rhetin = 1.0 μm and ρhetin = 0.01 kg m−3 (nhetin = 2.39 × 1012
m−3).
γ [–] p*in [kPa] pvdry [Pa] Tddry [°C] ΔTddry [°C] ΔTd1dry [°C] ηv [%]
Fig. 21. The dew point depression ΔTd versus the pressure loss ratio γ when
inlet heterogenous conditions are ρhetin = 0.01 kg m−3 and rhetin = 1.0 μm.
Table 5
Dehydration performance of supersonic separator B with different rhetin of inlet
foreign droplets when ρhetin = 0.01 kg m−3 and p*in = 250 kPa (γ = 0.6).
rhetin [μm] nhetin [m−3] pvdry [Pa] Tddry [°C] ΔTddry [°C] ηv
15
H. Ding, et al. Applied Thermal Engineering 171 (2020) 115074
Appendix A. Supplementary material axial-flow swirling generator and drainage structure for supersonic separators with
diversion cone, Chem. Eng. Res. Des. 133 (2018) 155–167.
[27] X. Liu, Z. Liu, Numerical investigation and improvement strategy of flow char-
Supplementary data to this article can be found online at https:// acteristics inside supersonic separator, Sep. Sci. Technol. 53 (6) (2018) 940–952.
doi.org/10.1016/j.applthermaleng.2020.115074. [28] W. Alnoush, M. Castier, Shortcut modeling of natural gas supersonic separation, J.
Nat. Gas Sci. Eng. 65 (2019) 284–300.
[29] C. Wen, X. Cao, Y. Yang, J. Zhang, Evaluation of natural gas dehydration in su-
References personic swirling separators applying the discrete particle method, Adv. Powder
Technol. 23 (2012) 228–233.
[1] M.I. Khan, M. Shahrestani, T. Hayat, A. Shakoor, M. Vahdati, Life cycle (well-to- [30] Y. Yang, C. Wen, CFD modeling of particle behavior in supersonic flows with strong
wheel) energy and environmental assessment of natural gas as transportation fuel in swirls for gas separation, Sep. Purif. Technol. 174 (2017) 22–28.
Pakistan, Appl. Energy 242 (2019) 1738–1752. [31] X. Liu, Z. Liu, Y. Li, Investigation on separation efficiency in supersonic separator
[2] S.F. Interlenghi, F.S. Raquel de Pádua, J.L. de Medeiros, O.Q.F. Araújo, Low- with gas-droplet flow based on DPM approach, Sep. Sci. Technol. 49 (2014)
emission offshore Gas-To-Wire from natural gas with carbon dioxide: supersonic 2603–2612.
separator conditioning and post-combustion decarbonation, Energ. Convers. [32] B. Prast, B. Lammers, M. Betting, CFD for supersonic gas processing, in: 5th Int Conf
Manage. 195 (2019) 1334–1349. on CFD in Process Industries CSIRO, Melbourne, Australia, 2006, pp. 1–6.
[3] X. Liu, Z. Liu, Investigation of the energy separation effect and flow mechanism [33] W. Jiang, J. Bian, A. Wu, S. Gao, P. Yin, D. Hou, Investigation of supersonic se-
inside a vortex tube, Appl. Therm. Eng. 67 (2014) 494–506. paration mechanism of CO2 in natural gas applying the Discrete Particle Method,
[4] C. Wen, N. Karvounis, J.H. Walther, Y. Yan, Y. Feng, Y. Yang, An efficient approach Chem. Eng. Process. 123 (2018) 272–279.
to separate CO2 using supersonic flows for carbon capture and storage, Appl. [34] Q. Ma, D. Hu, J. Jiang, Z. Qiu, A turbulent Eulerian multi-fluid model for homo-
Energy 238 (2019) 311–319. geneous nucleation of water vapour in transonic flow, Int. J. Comput. Fluid Dynam.
[5] P.B. Machado, J.G.M. Monteiro, J.L. de Medeiros, H.D. Epsom, O.Q.F. Araujo, 23 (2009) 221–231.
Supersonic separation in onshore natural gas dew point plant, J. Nat. Gas Sci. Eng. 6 [35] S.R. Shooshtari, A. Shahsavand, Reliable prediction of condensation rates for pur-
(2012) 43–49. ification of natural gas via supersonic separators, Sep. Purif. Technol. 116 (2013)
[6] S. Li, S. Deng, L. Zhao, R. Zhao, M. Lin, Y. Du, Y. Lian, Mathematical modeling and 458–470.
numerical investigation of carbon capture by adsorption: literature review and case [36] H. Ding, C. Wang, G. Wang, Analytic equations for the Wilson point in high pressure
study, Appl. Energy 221 (2018) 437–449. steam flow through a nozzle, Int. J. Heat Mass Transf. 91 (2015) 961–968.
[7] Y. Yang, J.H. Walther, Y. Yan, C. Wen, CFD modeling of condensation process of [37] M. Castier, Modeling and simulation of supersonic gas separations, J. Nat. Gas Sci.
water vapor in supersonic flows, Appl. Therm. Eng. 115 (2017) 1357–1362. Eng. 18 (2014) 304–311.
[8] Han J. Eriqitai, R. Duan, M. Wu, Performance of dual-throat supersonic separation [38] S. Dykas, M. Majkut, K. Smołka, M. Strozik, Comprehensive investigations into
device with porous wall structure, Chin. J. Chem. Eng. 22 (4) (2014) 370–382. thermal and flow phenomena occurring in the atmospheric air two-phase flow
[9] M. Haghighi, K.A. Hawboldt, M.A. Abdi, Supersonic gas separators: review of latest through nozzles, Int. J. Heat Mass Transf. 114 (2017) 1072–1085.
developments, J. Nat. Gas Sci. Eng. 27 (2015) 109–121. [39] S. Dykas, M. Majkut, K. Smołka, Strozik M Study of the wet steam flow in the blade
[10] X. Cao, J. Bian, Supersonic separation technology for natural gas processing: a re- tip rotor linear blade cascade, Int. J. Heat Mass Transf. 120 (2018) 9–17.
view, Chem. Eng. Process. 136 (2019) 138–151. [40] Q. Ma, D. Hu, G. He, Performance of inner-core supersonic gas separation device
[11] A.M. Teixeira, L.O. Arinelli, J.L. de Medeiros, O.Q.F. Araújo, Economic leverage with droplet enlargement method, Chin. J. Chem. Eng. 17 (6) (2009) 925–933.
affords post-combustion capture of 43% of carbon emissions: supersonic separators [41] H. Ding, C. Wang, C. Chen, Non-equilibrium condensation of water vapor in sonic
for methanol hydrate inhibitor recovery from raw natural gas and CO2 drying, J. nozzle, Appl. Therm. Eng. 71 (1) (2014) 324–334.
Env. Manag. 236 (2019) 534–550. [42] P.H. Niknam, H.R. Mortaheb, B. Mokhtarani, Dehydration of low-pressure gas using
[12] A. Karimi, M.A. Abdi, Selective dehydration of high-pressure natural gas using supersonic separation: experimental investigation and CFD analysis, J. Nat. Gas Sci.
supersonic nozzles, Chem. Eng. Process. 48 (2009) 560–568. Eng. 52 (2018) 202–214.
[13] C. Wen, X. Cao, Y. Yang, W. Li, An unconventional supersonic liquefied technology [43] S.R. Shooshtari, A. Shahsavand, Optimal operation of refrigeration oriented su-
for natural gas, Energy Educ. Sci. Technol. Part A Energy Sci. Res. 30 (2012) personic separators for natural gas dehydration via heterogeneous condensation,
651–660. Appl. Therm. Eng. 139 (2018) 76–86.
[14] M. Hammer, P.E. Wahl, R. Anantharaman, D. Berstad, K.Y. Lervåg, CO2 capture [44] S. Dykas, Numerical calculation of the steam condensing flow, Task Q 5 (2001)
from off-shore gas turbines using supersonic gas separation, Energy Procedia 63 51–135.
(2014) 243–252. [45] J. Bian, X. Cao, W. Yang, H. Du, P. Yin, Effects of external particles on the lique-
[15] P.H. Niknam, H.R. Mortaheb, B. Mokhtarani, Optimization of dehydration process faction property of natural gas in a Laval nozzle, Powder Technol. 339 (2018)
to improve stability and efficiency of supersonic separation, J. Natl. Gas Sci. Eng. 43 894–902.
(2017) 90–95. [46] Y. Yang, X. Zhu, Y. Yan, H. Ding, C. Wen, Performance of supersonic steam ejectors
[16] P.H. Niknam, H.R. Mortaheb, B. Mokhtarani, Effects of fluid type and pressure order considering the nonequilibrium condensation phenomenon for efficient energy
on performance of convergent-divergent nozzles: an efficiency model for supersonic utilisation, Appl. Energy 242 (2019) 157–167.
separation, Asia-Pacific J. Chem. Eng. 13 (2) (2018) 1932–2135. [47] M. Mccallum, R. Hunt, The flow of wet steam in a one-dimensional nozzle, Int. J.
[17] L.O. Arinelli, T.A.F. Trotta, A.M. Teixeira, J.L. de Medeiros, O.Q.F. Araújo, Offshore Numer. Meth. Eng. 44 (12) (1999) 1807–1821.
processing of CO2 rich natural gas with supersonic separator versus conventional [48] J. Marti, K. Mauersberger, A survey and new measurements of ice vapor pressure at
routes, J. Natl. Gas Sci. Eng. 46 (2017) 199–221. temperatures between 170 and 250 K, Geophys. Res. Lett. 20 (5) (1993) 363–366.
[18] J. Bian, X. Cao, W. Yang, M.A. Edem, P. Yin, W. Jiang, Supersonic liquefaction [49] X. Zhong, An improved generalized Watson equation for prediction of latent heat of
properties of natural gas in the Laval nozzle, Energy 159 (2018) 706–715. vaporization, Chem. Eng. Commun. 29 (1984) 257–269.
[19] M. Rezakazemi, N. Rahmanian, H. Jamil, S. Shirazian, Process simulation and [50] P.J. O’Rourke, Collective Drop Effects on Vaporizing Liquid Sprays, PhD thesis
evaluation of ethane recovery process using Aspen-HYSYS, Chem. Eng. Trans. 70 Princeton University, Princeton, New Jersey, 1981.
(2018) 961–966. [51] S. Dykas, W. Wróblewski, Two-fluid model for prediction of wet steam transonic
[20] G.V. Brigagão, L.O. Arinelli, J.L. de Medeiros, O.Q.F. Araújo, A new concept of air flow, Int. J. Heat Mass Transf. 60 (2013) 88–94.
pre-purification unit for cryogenic separation: low-pressure supersonic separator [52] H. Ounis, G. Ahmadi, J.B. McLaughlin, Brownian diffusion of submicrometer par-
coupled to finishing adsorption, Sep. Purif. Technol. 215 (2019) 173–189. ticles in the viscous sublayer, J. Colloid. Interf. Sci. 143 (1991) 266–277.
[21] J.L. de Medeiros, L.O. Arinelli, O.Q.F. Araújo, Speed of sound of multiphase and [53] C. Wang, P. Cao, H. Ding, X. Wang, D. Lin, Signal analysis of supersonic vapor
multi-reactive equilibrium streams: a numerical approach for natural gas applica- condensation in nozzle sensor using distributed transient pressure probes, IEEE
tions, J. Nat. Gas Sci. Eng. 46 (2017) 222–241. Trans. Instrum. Meas. 68 (4) (2018) 1053–1061.
[22] J.L. de Medeiros, L.O. Arinelli, A.M. Teixeira, O.Q.F. Araújo, Offshore Processing of [54] B.E. Wyslouzil, C.H. Heath, Binary condensation in a supersonic nozzle, J. Chem.
CO2-rich Natural Gas with Supersonic Separator: Multiphase Sound Speed, CO2 Phys. 113 (17) (2000) 7317–7329.
freeze-out and HYSYS Implementation, Springer Int. Publishing, 2019. [55] B. Wang, D. Xu, K. Chu, A. Yu, Numerical study of gas–solid flow in a cyclone
[23] M. Malyshkina, The procedure for investigation of the efficiency of purification of separator, Appl. Math. Model. 30 (11) (2006) 1326–1342.
natural gases in a supersonic separator, High Temp. 48 (2010) 244–250. [56] A.B. Wood, A Textbook of Sound, G. Bell and Sons Ltd., London, 1930.
[24] C. Wen, X. Cao, Y. Yang, J. Zhang, Swirling effects on the performance of supersonic [57] S.W. Kieffer, Sound speed in liquid-gas mixtures: water-air and water-steam, J.
separators for natural gas separation, Chem. Eng. Technol. 34 (9) (2011) Geophys. Res. 82 (20) (1977) 2895–2904.
1575–1580. [58] H. Ding, Y. Li, E. Lakzian, C. Wen, C. Wang, Entropy generation and exergy de-
[25] C. Wen, X. Cao, Y. Yang, Y. Feng, Prediction of mass flow rate in supersonic natural struction in condensing steam flow through turbine blade with surface roughness,
gas processing, Oil Gas Sci. Technol. 70 (2015) 1101–1109. Energ. Convers. Manage. 196 (2019) 1089–1104.
[26] Hu.D. WangY, Experimental and numerical investigation on the blade angle of
16