Professional Documents
Culture Documents
Dynamic Heat Transfer Analysis of Liquefied Natural Gas Ambient Air Vaporizer Under Frost Conditions
Dynamic Heat Transfer Analysis of Liquefied Natural Gas Ambient Air Vaporizer Under Frost Conditions
net/publication/308004230
CITATIONS READS
8 3,082
5 authors, including:
Hai-Chao Wang
Aalto University
33 PUBLICATIONS 480 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Liu Shanshan on 24 April 2018.
Research Paper
h i g h l i g h t s
a r t i c l e i n f o a b s t r a c t
Article history: Liquefied natural gas (LNG) ambient air vaporizer (AAV) is a cryogenic heat exchanger, which utilizes
Received 30 April 2016 ambient air to evaporate LNG in gas terminal stations. The moist air around the AAV fin tube may freeze
Revised 2 August 2016 during the long-time operation, leading to a worse heat transfer performance. Study on the dynamic heat
Accepted 4 September 2016
transfer performance of AAV is required for its optimal design and operation. This paper presents a cou-
Available online 7 September 2016
pled dynamic heat transfer model, which incorporates the cryogenic frost formation and flow boiling heat
transfer of LNG. The model was demonstrated in the heat transfer analysis of a typical star-fin type AAV.
Keywords:
The model results were validated by the test data of a commercial HAI848F type AAV. The dynamic heat
Liquefied natural gas (LNG)
Ambient air vaporizer (AAV)
transfer performance of AAV under frost conditions was also studied.
Dynamic Ó 2016 Elsevier Ltd. All rights reserved.
Heat transfer
Frost formation
1. Introduction LNG flows into the bottom of the fin tube, and natural gas flows
out of the fin tube. The outlet temperature of natural gas should
Natural gas can be liquefied at approximately 112 K under follow the requirement of the natural gas pipeline.
atmospheric pressure, but this can also happen by cooling down The heat transfer performance of AAV is easily affected by the
the natural gas with very high pressure in order to produce lique- atmospheric conditions and operation parameters, e.g. tempera-
fied natural gas (LNG). Liquefaction of natural gas makes the trans- ture and humidity of the ambient air. The temperature difference
portation much easier because it greatly reduces the volume of between ambient air and LNG in the tube is the driving force of
natural gas by 600 times. When LNG transportation is over, LNG heat transfer. Therefore, the ambient air temperature can affect
vaporizers, e.g. ambient air vaporizers (AAV) are designed for the the heat transfer coefficient, especially in cold areas, where the
vaporization of LNG in gas terminal stations. outlet temperature of the natural gas usually cannot satisfy the
AAV is widely used in small and medium gas terminal stations requirement of the natural gas pipelines. This means that the cli-
due to the low operation cost and high environmental sustainabil- mate conditions of different areas should be considered to design
ity. AAV has a group of parallel fin tubes, and the most common the AAV [1]. On the other hand, the cryogenic LNG may cause frost
AAV fin tube is with 8 fins or 12 fins in a group. AAV utilizes ambi- formation on the surface of the AAV fin tube. Frost layer makes the
ent air to heat the cryogenic LNG inside the fin tube. Generally, AAV ineffective because of the worse heat transfer coefficient, and
thus a second AAV has to be employed to let the first AAV defrost.
⇑ Corresponding author. This is why the operation time for a certain type of AAV needs to be
E-mail address: wljiao@163.com (W. Jiao). restricted. Otherwise, it will result in operating failures [2].
http://dx.doi.org/10.1016/j.applthermaleng.2016.09.016
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
1000 S. Liu et al. / Applied Thermal Engineering 110 (2017) 999–1006
Nomenclature
model is similar with the frost formation on the cold surface. How-
ever, the definite solution condition of the conservation equations
is quite different. The empirical correlations of the density and
thermal conductivity of frost layer on the cryogenic surface were
adopted to solve the equations.
According to the convective mass transfer, the water molecule
passes from the air into the frost layer via the gas-solid interface.
The mass growth of the frost layer equals to the convective mass
transfer rate.
dM f
¼ hm ðqv ;a qv ;f Þ ð2Þ
ds
Mf ¼ df qf ð2aÞ
where Mf represents the mass of the frost layer per square meter,
kg/m2; qf is the density of the frost layer, kg/m3; hm is the mass
transfer coefficient, which was obtained by the Chilton-Colburn
analogy [13].
hf
hm ¼ 2
ð2bÞ
qa cp;a Le3
where Le is the Lewis number, the ratio of the mass and diffusion
coefficient.
Thus, the mass conservation equation can be written as follows:
dðdf qf Þ
¼ hm ðqv ;a qv ;f Þ ð3Þ
ds
The total heat transfer from the ambient air to the frost layer
consists of the sensible heat and the latent heat. The energy con-
servation equation can be written as follows:
T sur T w
kf ¼ ha ðT a T sur Þ þ Isv mf ð4Þ
Fig. 1. The simplified heat transfer schematic diagram of AAV. df
In order to meet the definite solution condition of the mass and
energy conservation equations, the empirical correlations of the
density and thermal conductivity of frost formation on the cryo-
genic surface were introduced [14]. The frost layer is regarded as
a porous media composed of air and ice, so the thermal parameters
of the frost layer are affected by the structure.
1 f 1f
¼ þ ð5Þ
kf kmin kmax
where kmin and kmax is the minimum and maximum thermal con-
ductivity of the frost layer, and f is the density index of the frost
layer.
kmax ¼ ð1 wÞkice þ wka ð5aÞ
1 w
¼ ð1 wÞkice þ ð5bÞ
kmin ka
Fig. 2. The phase diagram of water. where w is the frost porosity, which can be calculated based on Eq.
(5d); kice is the thermal conductivity of the ice, and ka is the thermal
conductivity of the air; qf is the frost density and can be calculated
pressure of the water vapor. Although the temperature of the cryo- based on Eq. (5e).
genic surface of the AAV fin tube is much lower than the triple
point of water, the frost crystal may not form if the partial pressure w ¼ 1 0:71 exp ½0:228ðT sur 273:15Þ ð5dÞ
of water vapor was not in the proper range. Furthermore, the con-
tinuous accumulation of the frost crystal is essential to form the qf ¼ ð1 wÞqice þ wqa ð5eÞ
frost layer. The concentration gradient of the water vapor between
the ambient air and the moist air around the AAV fin tubes is the
2.2. LNG flow boiling inside the vertical smooth tube
driving force of the frost formation.
The mathematical model of the frost formation on the cryogenic
The vaporization of LNG inside the AAV fin tube is a convective
surface is based on the conservation laws of mass and energy. The
flow boiling of nonazeotropic mixture. The flow boiling involves
1002 S. Liu et al. / Applied Thermal Engineering 110 (2017) 999–1006
0:55
hnb ¼ 55P 0:12
c ðlog10 ðPc ÞÞ M0:5 q0:67 ð6bÞ
j>m No
0:9 !1
DT bp p q Yes
Fm ¼ 1 þ jY 1 X 1 j0:29 1 0:87exp
DT id 10 5
3 105 End
ð6cÞ
Fig. 3. Flowchart of calculation program for the AAV under frost formation.
where v is the vapor quality; Prl is the Prandtl number of the liquid;
G is the mass flux; DTbp is the temperature glide; DTid is the ideal
temperature difference; Y1 is the vapor phase composition of the 3.1. Thermal resistance of the frost layer
more volatile component; X1 is the liquid phase composition of
the more volatile component. Thermal resistance of the frost layer Rf will reduce the heat
The integrated heat transfer model is so complex that cannot be transfer coefficient of AAV as Eq. (1) shows, and it can be affected
solved analytically. Therefore, we use Matlab to solve the model by the operation time, atmospheric condition and the wall temper-
numerically and the calculation progress is developed to make ature of the AAV. The definition of Rf has the following form:
the solution concise and reliable. The tube of the AAV was assumed
df
to be one-dimensional due to the high slenderness ratio. The tube Rf ¼ ð7Þ
kf
segments and the iteration time were set based on the indepen-
dence tests. Fig. 3 illustrates the logic and flowchart of the calcula- where df is the thickness of the frost layer and kf is the thermal con-
tion. The numerical calculation consists of three parts divided by Tb ductivity of the frost. All the variables in Eq. (7) change with the
and Td. The local heat transfer coefficient hi of the single phase was operation time.
calculated using the single phase correlation, and hi of the mixture Fig. 4 shows the relationship between the thermal resistance of
phase was calculated using the Zou correlation. the frost layer and the operation time. It can be seen that the ther-
mal resistance increases with time. At the beginning of the frost
3. Results and discussions formation, the thermal resistance increases rapidly, and then
slowly as time goes on. This is because that the thickness of the
The frost formation increases the thermal resistance and frost layer grows fast at the beginning of the frost formation and
reduces the heat transfer coefficient of the AAV fin tube, and the then becomes slowly. When the thickness of the frost layer grows
heat transfer performance will get worse as the frost grows thicker. to a certain extent, the frost density begins to increase.
Based on the developed coupled model of cryogenic frost forma- Numerical calculations on different air temperatures were car-
tion and the coupled heat transfer, the thermal resistance of the ried out to check the influence of the air temperature on the ther-
frost layer and the coupled dynamic heat transfer performance of mal resistance of frost. Fig. 5 shows how the thermal resistance
AAV were investigated. changes with air temperature at different operation time. The ther-
S. Liu et al. / Applied Thermal Engineering 110 (2017) 999–1006 1003
Fig. 6. The change of frost thermal resistance with relative humidity. Ta = 293 K,
Tw = 200 K.
Fig. 4. The change of frost thermal resistance with time. Ta = 303 K, u = 50%,
Tw = 200 K.
Fig. 7. The change of frost thermal resistance with different wall temperature.
Ta = 293 K, u = 50%.
Fig. 5. The change of frost thermal resistance with different air temperature.
u = 50%, Tw = 200 K.
variation range of the thermal resistance with different wall tem-
perature is wide.
mal resistance firstly increases with air temperature and then It can be seen from the above analysis that the wall temperature
decreases, and the maximum value occurs at about 255–261 K is the major influencing factor, followed by the air temperature,
depending on the operation time. We named the corresponding and the relative humidity is the least influencing factor. Therefore,
air temperature of the maximum thermal resistance a ‘‘peak ther- the thermal resistance of frost on the AAV fin tube mainly varies
mal resistance temperature”. When the air temperature is lower or with the distance to the LNG inlet, because the wall temperature
higher than this range, the frost grows more slowly and heat trans- of the fin tube changes along with the distance to the LNG inlet.
fer coefficient is a little better. When the air temperature is lower,
the mass transfer of the water vapor is slow, forming a thinner 3.2. Dynamic heat transfer performance of AAV
frost layer. When the air temperature is higher, the back melting
and regrowth of the frost layer alternate quickly, leading to high Based on the proposed coupled heat transfer model of the AAV,
frost density and low thermal conductivity. numerical calculations and analysis on existing market AAV model
In order to investigate the effect of the air humidity on the ther- QQNG-2000/1.2 were conducted. The geometries and the operat-
mal resistance, numerical calculations were also carried out and ing parameters of this AAV model are listed in Table 1. The air tem-
the results are shown in Fig. 6. We found that the trend is similar perature is 283 K, and the relative humidity is 50%.
to Fig. 4, and the maximum thermal resistance value occurs when Fig. 8 shows that the thickness of the frost layer on the AAV fin
air humidity is about 30%. The thermal resistance ranges from 0.12 tube varies with the tube length and the operation time. No matter
to 0.18 (m2 K/W) when the air humidity changes from 5% to 100%, how long the AAV operates, the thickness of the frost layer near the
and this means that the air humidity has less influence on the ther- LNG inlet is the highest, and gradually become thinner along with
mal resistance compared to the air temperature. the tube length. The thickness declines almost linearly along the
Fig. 7 shows how the thermal resistance changes with different tube from 0 m to about 4 m, but increases sharply at about 4 m.
wall temperatures. The thermal resistance decreases almost lin- It is mainly because that the subcooled boiling begins at about
early as the wall temperature increases. There is no frost formation 4 m, which has a much higher heat transfer coefficient than the
when the wall temperature is above 270 K in this case. The single phase. Therefore the wall temperature declines rapidly at
1004 S. Liu et al. / Applied Thermal Engineering 110 (2017) 999–1006
Table 1
The geometries and operating parameters of the AAV.
Structure Nominal capacity Tin Operation pressure Number of tube Number of fin on each tube Di Do d H L
3
In parallel 2000 Nm /h 111 K 0.6 Mpa 10 12 28 mm 32 mm 4 mm 80 mm 48 m
Note: Tin is the inlet temperature of LNG; Di and Do is the inner diameter and external diameter of the fin tube; d is the thickness of the fin; H is the height of the fin; L is the
length of fin tube in series.
18 640
8 hours
16 560
Frost layer thickness (mm)
4 hours No frost
14
480
12
K(W/(m2·K))
2 hours 400
2 hours
10
1 hours 320 4 hours
8 1 hours 8 hours
240
6
160
4
80
2
0
0 0 4 8 12 16 20 24 28 32 36 40 44 48
0 4 8 12 16 20 24 28 32 36 40 44 48
Length(m)
Length(m)
Fig. 10. The coupled heat transfer coefficient of the AAV along with the tube length.
Fig. 8. The frost layer thickness along with the tube length at different operation
time.
310
300
16
290
14 4 hours 280
Outlet temperature(K)
8 hours 270
12 2 hours
260
hair (W/(m2·K))
1 hours
10 250
No frost
8 240
230 303K
6
220 293K
4 210 283K
200 273K
2
190
0 0 2 4 6 8 10 12 14 16 18 20 22 24
0 4 8 12 16 20 24 28 32 36 40 44 48 Operation time(h)
Length(m)
Fig. 11. The outlet temperature of natural gas related to the operation time at
Fig. 9. The convective heat transfer coefficient of the air along with the tube length. different air temperature.
S. Liu et al. / Applied Thermal Engineering 110 (2017) 999–1006 1005
Table 2
The geometries and operating parameters of the HAI848F type AAV.
Structure Nominal capacity Tin Operation pressure Number of tube Number of fin on each tube Di Do d H L
3
In parallel 524 Nm /h 135 K 1.6 Mpa 8 12 20 mm 24 mm 2 mm 82 mm 32 m
Inlet pressure transmitter Outlet pressure transmitter manufactured by Cryoquip company was numerically calculated.
LNG tank The numerical results were compared with the test data conducted
by the research department of Cryoquip [21]. The geometrical
parameters and the operating conditions of the numerical
calculation were the same with the tests by Cryoquip as Table 2
Gas pipeline shows. The test setup of the AAV vaporization system is shown
in Fig. 12.
The tests were conducted in winter and summer respectively.
The main test data were the outlet temperatures of natural gas
at different operation time. The inlet pressure, temperature, and
AAV the flow rate of LNG were kept constant during the test period.
Outlet temperature sensor
Inlet temperature sensor The temperature sensor is armoured platinum resistance with an
uncertainty of ±0.15 K. The air temperature and the relative
Psychrometer humidity were tested by the psychrometer with an uncertainty
Data monitor system of ±0.2 K and ±0.2% RH. All the pressure and temperature data
can be transmitted to the data monitor system.
Fig. 13 shows the comparisons between the numerical results
Fig. 12. Schematic diagram of test setup for LNG AAV system.
and the test data. It can be seen that the numerical results agree
well with the test data. The relative error was about 2.1% in
300
summer and 3.5% in winter, and the changing trends of Tout are
the same. The calculated Tout was a little lower than the test results
290
in summer, but a little higher in winter. In summer, the
Outlet temperature (K)
(3) Frost formation on the AAV greatly affects the coupled heat [7] S.P. Chen, F.S. Xie, S.T. Yao, Experiment study on unsteady frosting of cryogenic
vertical plate surface under natural convection conditions, Adv. Mater. Res.
transfer performance of the AAV, and the influence becomes
516 (2012) 24–29.
stronger as the operation time goes on. The coupled heat [8] V.A. Grigoryev, Yu.M. Pavlov, E.V. Ametistov, et al., Investigation on the growth
transfer coefficient along the tube length is heavily influ- rate of vapor bubbles in the boiling of cryogenic fluids, Heat Transf. Sov. Res. 7
enced by the frost formation. (1975) 133–140.
[9] Xiangdong Li, Wei Wei, Rongshun Wang, Yumei Shi, Numerical and
(4) According to the outlet temperature of natural gas, the effec- experimental investigation of heat transfer on heating surface during
tive operation time to switch the cycle of AAV can be subcooled boiling flow of liquid nitrogen, Int. J. Heat Mass Transf. 52 (2009)
obtained. For the same type of AAV, the effective operation 1510–1516.
[10] S. Nangle-Smith, J.S. Cotton, EHD-based load controllers for R134a convective
time becomes shorter when the air temperature is lower. boiling heat exchangers, Appl. Energy 134 (2014) 125–132.
[11] Dongsheng Chen, Yumei Shi, Experimental study on flow boiling heat transfer
of LNG in a vertical smooth tube, Cryogenics 57 (2013) 18–25.
[12] X. Zou, M.Q. Gong, G.F. Chen, Z.H. Sun, Y. Zhang, J.F. Wu, Experimental study on
Acknowledgements saturated flow boiling heat transfer of R170/R290 mixtures in a horizontal
tube, Int. J. Refrig. 33 (2010) 371–380.
One of the authors(Shanshan Liu) wishes to acknowledge the [13] T.H. Chilton, A.P. Colburn, Mass transfer coefficients prediction from data on
heat transfer and fluid friction, Ind. Eng. Chem. 26 (1934) 1183–1187.
technical advice and support from Pingyi Eden Security Engineer- [14] H. Auracher, Heat transfer in frost and snow, in: Proceedings of the 6th
ing Company and Linyi Yage Security Technology Company. International Heat Transfer Conference, Toronto, Canada 1978, pp. 25–30.
[15] K.E. Gungor, R.H.S. Winterton, Simplified general correlation for saturated flow
boiling and comparisons of correlations with data, Chem. Eng. Res. Des. 65
References
(1987) 148–156.
[16] D.S. Jung, M. McLinden, R. Rademacher, D. Didion, A study of flow boiling heat
[1] Robert E. Bernert Sr., New concepts for tropical cryogenic ambient air transfer with refrigerant mixtures, Int. J. Heat Mass Transf. 32 (1989) 1751–
vaporizers, in: The International Cryogenic Engineering Conference (ICEC 17), 1764.
Bournemouth, England, 1998, pp. 173–179. [17] Shanshan Liu, Wenling Jiao, Haichao Wang, Three-dimensional numerical
[2] R.E. Bernert Jr., W. Everett, R.E. Bernert Sr., Cryogenic ambient air vaporizers: analysis of the coupled heat transfer performance of LNG ambient air
frost growth, wind and seismic design for safety, Cryogenics 8 (1993) 789–793. vaporizer, Renew. Energy 87 (2015) 1105–1112.
[3] H.M. Jeong, Y.H. Lee, Natural convection heat transfer estimation from a [18] E.W. Lemmon, M.L. Huber, M.O. McLinden, Reference Fluid Thermodynamic
longitudinally fin vertical pipe using CFD, J. Mech. Sci. Technol. 23 (2009) and Transport Properties-REFPROP Version 8.0, NIST Standard Reference
1517–1527. Database, 2007.
[4] H.M. Jeong, H. Chung, Optimum design of vaporizer fin with Liquefied Natural [19] Ren Zepei, Convective Heat Transfer, Higher Education Press, Beijing, 1995.
Gas by numerical analysis, J. Mech. Sci. Technol. 20 (2006) 545–553. [20] Yan Mingqing, Gas Engineering Design Manual, China Building Industry Press,
[5] F. Gavelli, Computational fluid dynamics simulation of fog clouds due to Beijing, 2009.
ambient air vaporizers, J. Loss Prev. Process Ind. 23 (2010) 773–780. [21] Bryan Smith, Vaporizer ice built-up requires an analysis of switching issues for
[6] K.S. Lee, W.S. Kim, T.H. Lee, A one-dimensional model for frost formation on a ambient air units, Newslett. Cryogenics Ind. 8 (1996) 1–7.
cold flat surface, Heat Mass Transf. 40 (1997) 4359–4365.