Sie sind auf Seite 1von 10

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

net/publication/301294499

Thermal performance calculation and analysis of heat transfer tube in super


open rack vaporizer

Article  in  Applied Thermal Engineering · January 2016

CITATIONS READS

4 334

5 authors, including:

Jie Pan Ran Li


Xi'an Shiyou University Xi'an Jiaotong University
24 PUBLICATIONS   221 CITATIONS    15 PUBLICATIONS   62 CITATIONS   

SEE PROFILE SEE PROFILE

Tao Lv
Guangzhou Institue of Energy, Conversion
8 PUBLICATIONS   12 CITATIONS   

SEE PROFILE

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

Gas production from NGH View project

All content following this page was uploaded by Jie Pan on 22 May 2019.

The user has requested enhancement of the downloaded file.


Applied Thermal Engineering 93 (2016) 27–35

Contents lists available at ScienceDirect

Applied Thermal Engineering


j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g

Research Paper

Thermal performance calculation and analysis of heat transfer tube in


super open rack vaporizer
Jie Pan *, Ran Li, Tao Lv, Gang Wu, Zhian Deng
College of Petroleum Engineering, Xi’an Shiyou University, Xi’an 710065, Shaanxi Province, China

H I G H L I G H T S

• A distributed parameter model for SuperORV heat transfer tube was developed.
• Temperature and heat transfer coefficient profiles along tube length were estimated.
• The effects of operating parameters on thermal performance were discussed.
• The effects of inner fin or twisted tape insert on thermal performance were studied.

A R T I C L E I N F O A B S T R A C T

Article history: As one of the most widely-used liquefied natural gas (LNG) vaporizer, super open rack vaporizer (SuperORV)
Received 27 June 2015 consists of panel-shaped heat transfer tubes with duplex tube configuration. In this paper, an energy balance-
Accepted 17 September 2015 based distributed parameter model for predicting the thermal performance of SuperORV heat transfer
Available online 8 October 2015
tube was developed, where introduces numerous empirical correlations. The results from the model exhibit
a good agreement with the experimental data, which implies the calculation model is reliable. Based on
Keywords:
this, the heat transfer process of SuperORV heat transfer tube was simulated numerically, the bulk fluid
LNG
and metal temperature profiles along the heat transfer tube was obtained at the condition with ice for-
SuperORV
Heat transfer tube mation and heat transfer enhancement measures, the thermal performance of SuperORV heat transfer
Distributed parameter model tube was analyzed, and the effects of operating parameters and heat transfer enhancement measures
Thermal performance on the thermal performance were discussed. The results show that the operating parameters have im-
portant effects on the thermal performance of heat transfer tube, and the heat transfer enhancement
measures can improve the heat transfer performance markedly. The required minimum tube length while
both inner fin and twisted tape insert were applied is shortened by 60% compared with that without
heat transfer enhancement measures.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction key equipment is used to regasify LNG to meet the requirement of


fuel gas in industrial and domestic purposes. There are several types
Flexibility of supply and demand is essential for efficient oper- of mainly-used vaporization schemes in LNG receiving terminal, in-
ation of natural gas industry [1]. With the advantages consisting of cluding submerged-combustion vaporizer (SCV), intermediate fluid
high flexibility, and low storage and transportation cost, LNG links vaporizer (IFV), open-rack vaporizer (ORV) and SuperORV [3].
different markets together by allowing shifting volumes between As one of the most representative LNG vaporizers, convention-
regions, benefiting from differences in their supply and demand al ORV consists of panel-shaped heat transfer tubes, and seawater
balance [2], and has become the world’s fastest-growing primary is used for the heat source. Whereas, it is limited for the increase
energy commodity. The production and trade of LNG are increas- of vaporizing capacity and the decrease of seawater consumption
ingly active, and are becoming an important hotspot for oil and rate and installation space for conventional ORV due to the follow-
natural gas industry. ing reasons [4]: (1) the heat transfer efficiency reduces evidently
As important component of the LNG industry chain, numbers with increasing LNG flow rate; (2) it is difficult to form a falling water
of LNG receiving terminals have been commercially developed or film with uniform thickness over the external surfaces of the tubes
are under construction in many countries, where vaporizer as the when the rate of seawater decreases. Furthermore, the seawater
tends to form a thick ice layer on the outside surfaces of heat trans-
fer tubes due to direct heat transfer between seawater and cryogenic
* Corresponding author. Tel.: +86 29 88382932; fax: +86 29 88382932. LNG, and the formed thick ice results in a tremendous thermal re-
E-mail address: jackpan@xsyu.edu.cn (J. Pan). sistance that deteriorates the heat transfer efficiency greatly [5]. The

http://dx.doi.org/10.1016/j.applthermaleng.2015.09.047
1359-4311/© 2015 Elsevier Ltd. All rights reserved.
28 J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35

fins outside the heat transfer tubes are also choked with thick ice important effects on the vaporization efficiency and reliably of va-
layer until their configuration becomes almost invisible, which causes porizer. In this paper, a distributed parameter model for predicting
an effective decrease of heat transfer area [5]. thermal performance of heat transfer tube is developed, which in-
Therefore, a new type of open rack vaporizer (namely SuperORV) troduces numerous empirical correlations of heat transfer. Based
was jointly developed by Osaka Gas and Kobe Steel (Japan), which on the results from the model, a thermal performance analysis of
consists of panel-shaped heat transfer tubes similar to convention- superORV heat transfer tube was conducted, the structure design
al ORV except that the lower part of the tube panels has duplex tube for superORV heat transfer tubes was estimated, and the effects of
configuration. Compared with conventional ORV, the vaporizing ca- operating parameters such as LNG pressure and flow rate, seawater/
pacity of SuperORV was improved evidently and the seawater rate LNG flow rate ratio, LNG flow distribution ratio between annular
required for vaporization was reduced owing to the duplex tube con- channel, and inner tube and heat transfer enhancement measures
figuration. As well, the required installation space for SuperORV is on the heat transfer performance were discussed.
reduced and the operation cost was also cut down. The duplex tube
configuration of SuperORV heat transfer tube is characterized with 2. Physical and mathematical models
the annuli between outer and inner tubes, which enhances the heat
transfer between LNG and seawater available by suppressing icing 2.1. Description of physical problem
of seawater on the outside surfaces of heat transfer tubes, and the
length and thickness of ice layer decrease remarkably since the fins Fig. 1 illustrates the geometry of superORV heat transfer tube.
on the heat transfer tubes are maintained in good temperature As shown in Fig. 1, the lower half of the heat transfer tube is de-
condition. signed to be a duplex tube configuration differently from the upper
The heat transfer performance of superORV heat transfer tube section. The lower and upper parts of the heat transfer tube are called
was compared with that of conventional ORV heat transfer tube by “vaporizing section” and “heating section” respectively. The vapor-
Morimoto et al. [5], and it was found that SuperORV heat transfer izing section consists of an outer tube with petal-shaped internal
tube can contribute to not only lowering the cost for gas produc- fin and similar triangular external fin, and an inner tube with double
tion, but also saving energy. The results of pilot plant tests indicate twisted tape inserts. The heating section is only a circular tube with
that superORV increases LNG vaporizing rate per heat transfer tube petal-shaped internal fin and similar triangular external fin, where
3 to 5 times, reduces the operating cost, construction cost and in- cross-shaped twisted tape inserts are applied.
stallation space by 15%, 10%, and 40% respectively, and seawater was LNG with low temperature of −162 °C is distributed to different
reduced by approximately 15%. heat transfer tubes by the distribution header at the bottom of va-
With the aim to reveal and improve the performance of LNG va- porizer. The LNG flows upward inside heat transfer tubes while
porizer, numerous researchers respectively conducted a series of seawater flows downward along the outside surfaces of heat trans-
experimental study and numerical simulation. Jeong et al. [6] com- fer tubes. At vaporizing section, some of the cryogenic LNG flows
pared the characteristics of two types of longitudinal fin air- into the annular channel between outer tube and inner tube, and
heating vaporizers experimentally by analyzing the inlet–outlet is heated by external falling seawater film and vaporized rapidly,
enthalpy difference and the outlet fluid temperature at different and the other directly enters inner tube and is heated by absorb-
length scales and ambient conditions. A CFD analysis on longitu- ing the heat from outside NG until gasified to the saturated vapor.
dinally finned tube of air-heating vaporizer was conducted by Jeong The superheated NG from the annular channel and the saturated
et al. [7], the natural convection heat transfer for natural convec- NG flowing out of the inner tube are mixed and further heated in
tion outside the tube was estimated, and corresponding heat transfer heating section. After adequate heat absorption in heating section,
correlation was proposed. Jeong et al. [8] also carried out a numer- it turns to superheated vapor with ambient temperature, and at last
ical analysis of vaporizer for obtaining optimum design of assembles together in collection header at the top of vaporizer.
longitudinal fin air-heating vaporizer and the prediction frost deposit Outside and inside fins for the external tube, as well as a twisted
formation on vaporizer fin, and proposed the optimum vaporizer
geometry at the conditions of considering the frost thickness and
without frost deposit presence. Pu et al. [3] developed a thermal
model for intermediate fluid vaporizer based on the energy balance
among evaporator, condenser and thermolator, which considered
the mutual coupling and constraints fully. On this basis, the heat
transfer performance of vaporizer, and the effects of operating pa-
rameters, such as seawater temperature and flow rate, LNG pressure
and flow rate were also investigated. Himoto [9] analyzed the con-
ventional designs of submerged combustion vaporizer, including the
process and major factors impacting emissions, and systematical-
ly studied its advanced designs that have been newly developed to
reduce NOx and CO emissions in exhaust gases. Hisada and Sekiguchi
[10] established highly endurable structures through a variety of
tests or finite element method analysis to develop best-suited shapes
(such as sizes and internal/external surface shapes) of ORV heat
transfer tubes by using numerical analysis and thermal stress design.
Jin et al. [11] built a distributed parameter model to simulate the
LNG evaporating process in a SuperORV heat transfer tube, and ob-
tained some significant results by heat transfer performance analysis.
Complex heat transfer phenomena occur inside the heat trans-
fer tubes of superORV. Whereas, above review shows that previous
studies have mainly focused on air-heating vaporizer and few studies
reported the thermal performance of superORV heat transfer tube,
which is a key indicator for the thermal design of vaporizer and has Fig. 1. Geometry of superORV heat transfer tube.
J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35 29

tape insert in inner tube, are used to further improve the heat trans- ⎧  hsw 2
− h1sw
fer efficiency of heat transfer tube. ⎪msw = π ( do + 2b ) α sw 1
( Tsw1 − T01 )
⎪ Δ z
⎪ 2π k ( T01 − T11 )
⎪π ( do + 2b ) α sw ( Tsw − T0 ) =
1 1 1
2.2. Mathematical model ⎨ ln (1 + 2b do ) (2)

Based on the steady-state heat transfer theory, a distributed pa- ⎪π ( do + 2b ) α 1
sw ( T 1
sw − T 1
0 ) = π diα i1 ( Tl1 − Tg1 )
⎪ h2 − h1g
rameter model was developed for the vaporizing section under icing ⎪mg g = π diα i1 ( Tl1 − Tg1 )
condition, on the assumption that the flow pressure drop along the ⎩ Δz
heat transfer tube, the axial and circumferential heat transfer, the
At no icing condition, the heat transfer equations of vaporizing
radial thermal resistance of internal and external tube wall, and the
section, with 5 independent unknown variables of Δz , T11, T21, hsw2
,
effects of formed ice layer on heat transfer area are ignored. Based
hg2 , were given by:
on above assumption, Fig. 2 shows the heat flow transfer process
of vaporizing section with ice layer presence. ⎧  hsw 2
− h1sw
This model consists of some discretized equations derived from ⎪msw = π d1,oα sw 1
( Tsw1 − T11 )
⎪ Δ z
the energy balance of heat transfer tube, which illustrates that: (1)
⎪π d1,oα sw ( Tsw − T0 ) = π d1,iα1,i ( T1 − Tg )
1 1 1 1 1 1
the increase of seawater enthalpy, the convective heat transfer rate ⎪⎪
h − hg
2 1
from the seawater to the external surface of ice layer, the heat trans- ⎨mg g = π d1,iα11,i ( T11 − Tg1 ) − π d2,oα 21,o ( Tg1 − T21 ) (3)
fer rate through the ice layer by thermal conduction, and the ⎪ Δz
convective heat transfer rate from the internal surface of outer tube ⎪π d2,oα 21,o ( Tg1 − T21 ) = π d2,iα 21,i ( T21 − Tl1 )
to the NG in annular channel are equal; (2) the enthalpy changes ⎪
⎪  hl − hl
2 1

⎪⎩ml Δz = π d2,iα 2,i ( T2 − Tl )


1 1 1
of the NG flowing in annular channel and the LNG inside inner tube
are respectively equal to the sum of the heat transfer rates through
its boundaries; (3) the heat flow transferred from NG to inner tube and the heat transfer equations of heating section without ice layer
wall is equal to that from inner tube wall to LNG. The proposed presence were obtained by the same way:
discretized equations are expressed as follows:
⎧  hsw 2
− h1sw
⎧  hsw 2
− h1sw ⎪msw = π doα sw 1
( Tsw1 − T11 )
= π ( d1,o + 2b ) α sw
1
( Tsw1 − T01 ) ⎪⎪ Δ z
⎪msw Δz ⎨π doα sw ( Tsw − T0 ) = π diα i ( T1 − Tg )
1 1 1 1 1 1
⎪ (4)
⎪ 2π k ( T01 − T11 ) ⎪ h − hg
2 1
⎪π ( d1,o + 2b ) α sw ( Tsw − T0 ) = ⎪m = π diα i1 ( T11 − Tg1 )
1 1 1
 g
⎪ ln (1 + 2b d1,o ) ⎪⎩ g Δz
⎪⎪π ( d1,o + 2b ) α sw

1
( Tsw − T0 ) = π d1,iα11,i ( T11 − Tg1 )
1 1
(1) where only 3 independent unknown variables exist, including Δz ,
⎪  hg − hg
2 1

⎪m g Δz = π d1,iα1,i ( T1 − Tg ) − π d2,oα 2,o ( Tg − T2 )


1 1 1 1 1 1 T11 , hsw
2
.

⎪π d2,oα 2,o ( Tg − T2 ) = π d2,iα 2,i ( T2 − Tl )
1 1 1 1 1 1
2.3. Heat transfer correlations

 hl − hl = π d2,iα 21,i ( T21 − Tl1 )
2 1
⎪m
⎪⎩ l Δz Diverse heat transfer mechanisms exist in different parts of
SuperORV heat transfer tube, which are the bases for thermal cal-
An assumption different from Ref. [11] was applied in the present culation. In the present model, the heat transfer tube was divided
model, where the external surface temperature of ice layer was con- into different heat transfer regimes through corresponding crite-
sidered to be a known parameter slightly smaller than the freezing ria, including pure liquid convection zone, subcooled boiling zone,
point of seawater, and the ice thickness was treated as an unknown saturated boiling zone, mist flow evaporation zone and super-
variable. There are 6 independent unknown variables in the above- heated vapor convection zone, and a series of applicable empirical
mentioned equations, containing Δz , b, T11, T21, hsw
2
, hg2 . correlations were introduced to evaluate the heat transfer in various
Considering the ice formation, the heat transfer equations of zones.
heating section were obtained in the similar manner, which has 4
independent unknown variables including Δz , b, T11, hsw 2
:
2.3.1. Convective heat transfer for single-phase fluid
A trial calculation shows that the Reynolds numbers of the
subcooled LNG in inner tube and the superheated NG in both annular
channel and heating section were more than 104, thereby the flow
profiles all belonged to turbulent flow according to corresponding
criterion. Therefore, the classic Dittuse–Boelter correlation was
applied to calculate the heat transfer coefficients of above-mentioned
single-phase fluid, and the formula is expressed by [12]:

Nu = 0.023Re0.8 Pr n (5)

where n = 0.4 for heating the fluid and n = 0.3 for cooling the fluid,
and the scope of application is l/d > 60, Re > 104, 0.7 < Pr < 120.

2.3.2. Heat transfer for subcooled flow boiling


For partial subcooled flow boiling, the total heat flux was assumed
to be the sum of contributions due to single-phase forced
convection qspl and nucleate boiling qsnb, which can be determined
Fig. 2. Heat flow of vaporizing section with ice layer presence. by Rohsenow method [12]. In this method, conventional
30 J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35

single-phase correlation was used to predict the convective ⎛ ρ k μ Pr ⎞ x + (1 − x )( ρ v ρl )( kv kl )


contribution, and the nucleate boiling contribution was Γ ⎜ x, l , l , v , v ⎟ =
⎝ ρ v kv μl Prl ⎠ x + (1 − x ) ( ρ v ρl )( μv μl )
computed by: (15)
0.4
⎡ x + (1 − x )( ρ v ρl ) ⎤
1 1
⎢ ⎥
⎡ g ( ρl − ρ v ) ⎤
2 −
0.33
⎡ c (T − T ) ⎤ 0.33 ⎣ x + (1 − x )( ρ v ρl )( Prv Prl ) ⎦
qsnb = μl ilv ⎢ ⎥ Pr 1.7 ⎢ pl w sat ⎥ (6)
⎣ σ ⎦ ⎣ C sf ilv ⎦
The recommended value of x is 0.95 according to Ref. [12].
The detailed definition for the correlation can be found in Ref.
[12]. 2.3.5. Heat transfer for convective condensation
In two-phase region, the heat transfer between the outside surface
of inner tube and the NG in annular channel was considered to be
2.3.3. Heat transfer for saturated flow boiling convective condensation. A correlation for downflow internal con-
The Chen correlation was used to calculate the heat transfer co- vective condensation from Ref. [12] was used to calculate the heat
efficients for saturated convective boiling, which is based on linear transfer in saturated boiling zone, which is cast in the form:
superposition presupposition of the microscopic (nucleate boiling)
0.5
and microscopic (bulk convective) contributions, and the formula ⎛ 20 1 ⎞
α = 0.028 ⎛⎜ l ⎞⎟ Rel0.9 Prl0.5 ⎜ 1 +
k
is expressed by: + ⎟ (16)
⎝d⎠ ⎝ Xtt Xtt2 ⎠
α tp = α mic + α mac (7)
For mist flow evaporation zone, the heat transfer between the
outside surface of inner tube and the NG in annular channel was
In Chen correlation, the microscopic contribution is evaluated
assumed to be convective condensation with annular flow, and the
as:
local heat transfer coefficient correlation was proposed as [12]:
α mac = α l F ( Xtt ) (8) 0.5
⎛ 1.376 + 8 Xtt1.655 ⎞
α = 0.0274 ⎛⎜ l ⎞⎟ Rel0.6792Prl x0.2208 ⎜
k
⎟ (17)
where ⎝ ⎠
d ⎝ Xtt2 ⎠

⎧1 for Xtt−1 ≤ 0.1 The detailed definition for the correlation can be found in

F ( Xtt ) = ⎨ ⎛ 1 ⎞
0.736 (9) Ref. [12].
⎪2.35 ⎜ 0.213 + ⎟ for Xtt−1 > 0.1
⎩ ⎝ Xtt ⎠
2.3.6. Heat transfer for seawater falling film flow
and the single-phase coefficient for liquid alone hl is calculated using The thickness of ice layer on the outside surfaces of heat trans-
the Dittuse–Boelter equation, fer tube is seen to be invariable at steady state conditions. Thereby,
seawater only released its explicit heat to the NG in heat transfer
α l = 0.023⎛⎜ l ⎞⎟ Rel0.8 Prl0.4
k
(10) tube by convective heat transfer. The Wilke formula [13] was applied
⎝d⎠
to the convective heat transfer of seawater side, which are given
The microscopic contribution to the overall heat transfer coef- by
ficient is determined by applying the following correction:
α * = 0.00681Re0.43 Pr 0.344 (18)
1
⎡ k c ρ 0.79 0.45
⎤ 0.49 2 where the dimensionless heat transfer coefficient α * is defined by
hmic = 0.00122 ⎢ 0.5 0.29 0.24 0.24 ⎥ [Tw − Tsat ( pl )] [ psat ( Tw ) − pl ] S
l pl l 0.24 0.75

⎣ σ μ l hlv ρ v ⎦
α * = ( h λ ) (ν 2 g )
13
(11)
where
The applicable range of the above formula is Re > 3200,
S ( Retp ) = (1 + 2.56 × 10−6 Re1tp.17 )
−1
(12) 5 < Pr < 210.

Retp = Rel [ F ( Xtt )]


1.25
(13) 2.4. Enhanced heat transfer calculation

Three heat transfer enhancement measures, including similar tri-


2.3.4. Heat transfer for mist flow evaporation angular fin, petal-shaped fin, and twisted tape insert are applied
At low to moderate heat flux and wall superheat levels, the crit- in SuperORV heat transfer tubes, as introduced in Ref. [4,5]. The
ical heat flux (CHF) transition corresponds to liquid film dryout and similar triangular fins and petal-shaped fins are respectively applied
is a transition to mist flow evaporation. The mist flow evaporation on the external and internal surface of both the outer tube of va-
regime is also referred to as liquid deficient region and the regime porizing section and the heating section part, which can enhance
of dispersed flow heat transfer. The dispersed mist flow behaves like the heat transfer between bulk fluid and tube wall by extending heat
single-phase flow of a pseudo-fluid with appropriately defined mean transfer surface and increasing turbulence intensity. The cross-
properties of vapor and liquid, thereby the homogeneous flow model shaped twisted tape insert is applied in the inner tube of vaporizing
was introduced in the Dittuse–Boelter correlation to evaluate the section to enhance the convective heat transfer because vortex flow
mist flow evaporation heat transfer [12]: induces the reduction of thermal boundary layer thickness and the
increase of secondary flow intensity.
⎛ ρ k μ Pr ⎞ Since the actual heat transfer coefficient of seawater side is far
Nuv = 0.023 Re0v.8 Prv04 Γ ⎜ x, l , l , v , v ⎟ (14)
⎝ ρ v kv μl Prl ⎠ lower than calculated value, the enhancement effect of the similar
triangular fin outside the heat transfer tube is ignored in this paper,
where the correction factor Γ represents the average property of and the heat transfer coefficient for numerical calculation is 85%
the two-phase pseudo-fluid: of the results computed from Wilke correlation, in the light of the
J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35 31

comparison between the calculated tube length and the experi- Table 2
mental data from Ref. [10]. The enhancement mechanism of twisted Basic technical parameters for numerical simulation.

tape insert on heat transfer can be attributed to the following three Parameter Value
aspects: smaller hydraulic diameter, stronger secondary flow and LNG mass flow rate (kg/s) 0.0972
good thermal contact between tape and tube wall [14]. In the present Inlet temperature of LNG (K) 130
model, a twisted tape with a helix angle of 45° was chosen for inner Seawater outlet temperature (K) 277.65
tube, which can improve the Nusselt number to 1.8 times for tur- Seawater pressure (MPa) 0.1
Required outlet temperature of NG (K) 275
bulent flow according to Ref. [15]. The petal-shaped internal fin in Seawater freezing point (K) 271.25
outer tube was equivalent to rectangular straight fin with the fin Thermal conductivity of ice (W m−1 K−1) 2.22
height of 3.0 mm and the fin base thickness of 2.2 mm, and its en-
hancement effects can be evaluated by fin efficiency and increased
surface area. The fin efficiency of rectangular straight fin is calcu-
temperature was supposed to be the freezing point of seawater
lated by
which is generally considered to be 271.25 K (−1.9 °C). The thermal
tanh ( mH ′ ) conductivity of ice was postulated to be 2.22 W m−1 K−1, and the at-
ηf = (19) mospheric pressure of 0.1 MPa was adopted as the outlet pressure
mH ′
of seawater.
and
4. Results and discussion
2α ⎛H + δ ⎞
mH ′ = ⎜ ⎟ (20)
kδ ⎝ 2⎠ 4.1. Model validation and heat transfer performance analysis

3. Solution procedure To assess the predictive accuracy of the proposed model, a nu-
merical simulation for the thermal performance of conventional ORV
A FORTRAN code based on above-mentioned distributed pa- heat transfer tube was conducted. Fig. 4 shows the comparison
rameter models was developed to calculate the thermal performance between the calculation results and the data from the experi-
of SuperORV heat transfer tube numerically. In the calculation, ments in Ref. [10]. Most of the calculation results are in good
methane as the main composition was used to replace the LNG and agreement with the experimental data, which implies that the nu-
NG, and its thermal properties of methane were computed with a merical simulation with this model is reliable. As illustrated in Fig. 4,
code in the REFPROP package developed by the National Institute
of Standards and Technology (US) [16]. Because the salt content of
seawater is very little, a subroutine for thermodynamic properties
of water and steam based on IAPWS-IF97 was applied in the thermal
properties calculation of seawater.
In the iterative solution of models, the SuperORV heat transfer
tube was divided into different heat transfer elements along the flow
direction based on constant enthalpy difference. For each element,
the outlet parameters can be calculated from the inlet parameters
according to the energy conservation equations, and numerous heat
transfer correlations and thermal properties data. Then, they were
applied in the calculation of next element as the inlet parameters.
At last, the data for all elements can be obtained by the model
solution.
The thermal performance calculation and analysis of SuperORV
heat transfer tube was carried out for the following two cases, dif-
fering by whether these form ice layers on the outside surfaces of
outer finned tube. The superheated NG flowing out of the annular
channel and the saturated NG from the inner tube converge in the
heating section. The temperature of the mixed NG at the inlet of
the heating section was derived from its enthalpy which was pos-
tulated to be the sum of those of the NG out of the inner tube and
the annular channel.
The basic geometrical and technical parameters available for the
numerical calculation of heat transfer tube are respectively sum-
marized in Table 1 and Table 2, and the calculation procedure is
illustrated in Fig. 3. In the present calculation, the ice’s surface

Table 1
Basic geometrical parameters of SuperORV heat transfer tube.

Parameter Value

Outside diameter of outer tube (mm) 36


Inside diameter of outer tube (mm) 22
Outside diameter of inner tube (mm) 18
Inside diameter of inner tube (mm) 16
Helix angle of twisted tape insert 45°
Fig. 3. Logical diagram of the calculation program.
32 J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35

Fig. 4. Temperature profiles in conventional ORV heat transfer tube. Fig. 6. Heat transfer coefficient profiles in SuperORV heat transfer tube.

an obvious reduction of finned tube temperature is found to exist


measures) profiles in SuperORV heat transfer tube at the pressure
in two-phase zone on account of the heat transfer enhancement re-
of 4 MPa, where the operating parameters and boundary condi-
sulting from flow boiling. There is no similar change trend in practical
tions for numerical calculation are also given. In Fig. 5, at the lower
case owing to the existence of the heat transfer difference between
part of the heat transfer tube, the monotonous temperature rise of
LNG and methane in two-phase zone.
the low temperature LNG in inner tube and annular channel with
In the lower part of heat transfer tube, the low temperature LNG
increasing tube length derived from continual decalescence, and the
temperature and the finned tube temperature increase monoto-
temperature curves of external finned tube and inner tube show a
nously with increasing tube length through continual decalescence.
similar change trend. At the same location, the heat transfer coef-
While the tube length reaches 1.36 m, the flow enters two-phase
ficients for the seawater side and the inside of inner tube are almost
zone and the tube wall temperatures have a sharp decrease due to
constant, and these for the inside of finned tube and the outside
the heat transfer enhancement of two-phase boiling. Then, the LNG
of inner tube have a slight increase, as shown in Fig. 6.
temperature is maintained at saturation value of 177.2 K, as the tube
While the tube length reaches 0.99 m, the flow in annular channel
wall temperature experience an obvious jump with respect to the
enters two-phase zone, the heat transfer coefficients for the inside
abrupt augment of convective heat transfer coefficient in different
of finned tube and the outside of inner tube are augmented evi-
heat transfer regimes. At a tube length of 2.36 m, the finned tube
dently due to the heat transfer enhancement of two-phase boiling,
temperature has an abrupt rise by reason of liquid film dryout. While
and commensurately, the wall temperature of inner tube experi-
the tube length exceeds 2.53 m, the finned tube temperature in-
ences an obvious jump. In two-phase zone, the LNG is maintained
creases speedily once again with the temperature increase of the
at the saturation temperature of 177.2 K, and the inner tube shows
superheated NG in tube.
a similar trend with a constant temperature slightly lower than the
Figs. 5 and 6 respectively present the temperature and heat trans-
saturation value. Whereas, the wall temperatures of external finned
fer coefficient (considering the effects of heat transfer enhancement
tube decrease sharply and then have an abrupt augment at tube
length of 1.62 m accompanying obvious jump of heat transfer co-
efficient, by reason of liquid film dryout.
While the tube length exceeds 1.69 m, the superheated NG and
metal temperatures of the heat transfer tube increase speedily once
again along the flow direction, and the heat transfer coefficients for
the inside of finned tube and the outside of inner tube are reduced
commensurately. The wall temperature increase of external finned
tube is slowed down over the tube length of 2.40 m because no ice
layer forms on the outside surface of finned tube accompanied by
a low and almost constant heat transfer coefficient. Similar tem-
perature trend of the LNG in annular channel and the inner tube
appears at a larger tube length.
Also, constant LNG temperature and obvious jump of the inner
tube temperature occur in inner tube at the tube length from 3.02 m
to 3.95 m and the heat transfer coefficient between LNG and inner
tube is augmented markedly owing to two-phase flow boiling, ac-
companied by almost invariable temperature of the finned tube and
the NG in annuli. At the tube length of 3.95 m, the temperatures
of finned tube and NG experience obvious jumps because the mixing
process of superheated NG flowing out of the annular channel and
the saturated NG from the inner tube at the interface between the
Fig. 5. Temperature profiles in SuperORV heat transfer tube. vaporizing and heating sections were ignored. Similar jump exists
J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35 33

Fig. 8. Thermal performance of heat transfer tube at different LNG pressure.


Fig. 7. Comparison of ice thickness between SuperORV and conventional ORV.

Fig. 9 illustrates the required minimum length of LNG vaporiz-


ing section and heating section, and the length and thickness of ice
in the heat transfer coefficient of the heating section. While the tube layer formed outside the finned tube at different seawater/LNG flow
length exceeds 3.95 m, the mixed NG flows into the heating section, rate ratio. In Fig. 9, the ice thickness and the length of ice layer are
the heat transfer coefficient of NG almost keeps constant, and the reduced monotonously with increasing seawater/LNG ratio, and the
temperature of finned tube increases gradually with increasing tube required length of LNG vaporizing section shows a similar trend.
length because of continuous heat absorption, at last near the sea- These are because the heat transfer between seawater and outside
water temperature. The required minimum tube length of vaporizing surface of finned tube is enhanced evidently. In contrast, the re-
section and heating section are 3.95 m and 2.596 m respectively. As quired length of LNG heating section has an obvious increase with
shown in Fig. 5, the temperature variation of seawater is about the augment of seawater/LNG ratio. The reason is that low heat trans-
3.4 K, which results in an almost invariable heat transfer coeffi- fer rate results from low seawater temperature, as illustrated in Fig. 9,
cient in Fig. 6. because the larger the seawater/LNG ratio is, the lower the average
The thickness distribution curves of ice layer formed outside the temperature of seawater outside the finned tube becomes.
heat transfer tube of SuperORV and conventional ORV are given in Fig. 10 shows the required minimum length of LNG vaporizing
Fig. 7. It shows that the length and average thickness of ice layer section and heating section, and the length and thickness of ice layer
on conventional ORV are 2.88 m and 3.15 mm respectively. Com- formed outside the finned tube at different LNG flow distribution
pared with conventional ORV, SuperORV has a shorter length of ice ratio between annular channel and inner tube. In Fig. 10, the ice
layer of 2.40 m and a smaller ice thickness of 2.89 mm, even with thickness and the length of ice layer rise monotonously with in-
a larger LNG flow rate, which indicates the heat transfer enhance- creasing LNG flow distribution ratio in virtue of lower temperature
ment effect of duplex tube configuration. of external finned tube derived from the heat transfer enhance-
ment between external finned tube and the LNG/NG in annular
4.2. Effects of operating parameters channel. The required length of LNG vaporizing section and heating
section shows an opposite trend due to the same reason.
Fig. 8 shows the required minimum length of LNG vaporizing
section and heating section, and the length and thickness of ice layer
formed outside the finned tube at different LNG pressure. As shown
in Fig. 8, the ice thickness is reduced and the length of ice layer rises
monotonously with increasing pressure since the two-phase zone
in annular channel becomes shorter which weakens the heat trans-
fer inside finned tube at icing region. However, the required
minimum tube length shows a nonmonotonic change trend with
the increased LNG pressure. The required length of vaporizing section
rises from 4.02 to 4.21 m while the LNG pressure is augmented from
2 to 2.5 MPa, then decreases continually with increasing pressure,
and is reduced to 3.95 m at pressure of 4.0 MPa. This is because the
enthalpy of saturated vapor is reduced with increasing pressure al-
though the heat transfer is weakened at the same time owing to
the shortened two-phase boiling zone. In Fig. 8, the required length
of heating section has a little decrease from 2.52 to 2.45 m with the
increase of LNG pressure from 2.0 to 3.0 MPa, and then increases
to 2.60 m when LNG pressure falls to 4.0 MPa. It depends on the en-
thalpy of the mixed NG at the interface between the vaporizing and
heating sections and the heat transfer in heating section which is
enhanced with the augment of pressure. Fig. 9. Thermal performance of heat transfer tube at different seawater/LNG ratio.
34 J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35

all used, because the above-mentioned two opposite influences may


cancel out. As indicated in Fig. 11, the required minimum tube length
can be reduced by 60% when both inner fin and twisted tape insert
were applied in SuperORV heat transfer tube.
All in all, the heat transfer enhancement measures can improve
the heat transfer performance markedly, although the enhanced heat
transfer will induce more serious icing which has negative effect
on heat transfer. In addition, the results show that the heating section
is more susceptible to the heat transfer enhancement measures than
the vaporizing section.

5. Conclusions

A distributed parameter model was developed in this paper to


explore the LNG evaporating process in the new-type heat trans-
fer tube. In terms of diverse heat transfer mechanisms, the heat
transfer tube was divided into different heat transfer regions by cor-
responding criteria, and appropriate correlations were introduced
to evaluate the heat transfer in various regions.
Fig. 10. Thermal performance of heat transfer tube at different LNG flow distribu- The model was validated by experimental data from Ref. [10] and
tion ratio. the comparison shows a good agreement, which implies that the
distributed parameter model is reliable. Based on this, the heat trans-
fer process of SuperORV heat transfer tube was simulated
4.3. Effects of heat transfer enhancement measures numerically, the temperature and heat transfer coefficient temper-
ature profiles along the heat transfer tube was obtained at the
Fig. 11 illustrates the required minimum length of LNG vapor- condition with ice formation and heat transfer enhancement mea-
izing section and heating section, and the length and thickness of sures, the thermal performance of SuperORV heat transfer tube was
ice layer formed outside the finned tube with different heat trans- analyzed, and the effects of LNG pressure and flow rate, seawater/
fer enhancement methods applied. In Fig. 11, the required length LNG flow rate ratio, LNG flow distribution and heat transfer
of LNG vaporizing section and heating section while there is no heat enhancement measures on the thermal performance including the
transfer enhancement measures apply, are far larger than those for length of vaporizing section and heating section, the length and thick-
the heat transfer tube with inner fin or twisted tape insert, and the ness of ice layer formed outside the finned tube were discussed.
length of formed ice layer is augmented at heat transfer enhance- The results show that the operating parameters have impor-
ment conditions. These indicators are minimums when both the tant effects on the thermal performance of SuperORV heat transfer
inner fin and twisted tape insert are all used to enhance heat trans- tube. The required minimum tube length of vaporizing section is
fer. The change trend of ice thickness is, nevertheless, more complex reduced with increasing seawater/LNG ratio and LNG flow distri-
while different heat transfer enhancement methods are applied. The bution ratio, and that of heating section rises with the increase of
thickness of formed ice layer has a little decrease when the heat seawater/LNG ratio and decreases with the augment of LNG flow
transfer inside outer tube is improved by fin structures. However, distribution ratio. However, the required tube length shows a
it is increased when twisted tape insert is applied. The main reason nonmonotonic change trend with the increased LNG pressure. The
is that the heat transfer enhancement in inner tube resulting from ice thickness decreases with increased LNG pressure and seawater/
twisted tape insert causes the decrease of finned tube tempera- LNG flow rate ratio, and increases with the augment of LNG flow
ture, which benefits ice layer formation. The increase of ice thickness distribution ratio. The length of ice layer rises with increasing LNG
is weakened when both the inner fin and twisted tape insert are pressure and LNG flow distribution ratio, and is reduced with the
increase of seawater/LNG ratio. The required seawater is sensitive
to the change of seawater/LNG ratio.
The heat transfer enhancement measures can improve the heat
transfer performance markedly, and the required tube length can
be reduced by 60% although the enhanced heat transfer will induce
more serious icing which has negative effect on heat transfer.

Acknowledgements

This study was supported by the National Natural Science Foun-


dation of China (Grant No. 51304160) and Natural Science Basic
Research Plan in Shaanxi Province of China (Program No.
2014JQ7274).

Nomenclature

A heat transfer area, m2


b thickness of ice, m
Csf constant referring liquid–surface combination
cp specific heat at constant pressure, J kg−1 K−1
Fig. 11. Thermal performance of heat transfer tube with different heat transfer en- d tube diameter or hydraulic diameter, m
hancement measures. g acceleration of gravity, m s−2
J. Pan et al./Applied Thermal Engineering 93 (2016) 27–35 35

H fin height, mm mac macroscopic (bulk convective) contribution


h specific enthalpy, J kg−1 o outside surface
hlv enthalpy difference between saturated vapor and satu- sw seawater
rated liquid, J kg−1 snb nucleate boiling
i latent heat, J/kg−1 K−1 sat saturation state
k thermal conductivity, W m−1 K−1 tp two-phase
m  mass flow rate, kg s−1 v vapor
Nu Nusselt number w wall
ηf fin efficiency
Pr Prandtl number References
p Pressure, Pa
Q heat flow, W [1] S. Kumar, H.T. Kwon, K.H. Choi, et al., Current status and future projections of
q heat flux, , W m−2 LNG demand and supplies: a global prospective, Energy Policy 39 (7) (2011)
Re Reynolds number 4097–4104.
[2] S. Cornot-Gandolphe. LNG cost reductions and flexibility in LNG trade add to
S suppression factor security of gas supply. IEA (2005): Energy Prices and Taxes, Quarterly Statistics.
T temperature, K First Quarter, 2005.
Xtt Martinelli number [3] L. Pu, Z. Qu, Y. Bai, et al., Thermal performance analysis of intermediate fluid
vaporizer for liquefied natural gas, Appl. Therm. Eng. 65 (1) (2014) 564–574.
x thermal equilibrium quality [4] K. Yamazaki, T. Shimokawatoko, Y. Yamasaki, et al., Development of a New Type
Δz element length, m of Open Rack LNG Vaporizer, Osaka Gas Co. Ltd. & Kobe Steel, Ltd., 1998.
[5] N. Morimoto, S. Yamamoto, Y. Yamasaki, et al., Development and practical
application of a high performance open-rack LNG vaporizer, Osaka Gas Co. Ltd.,
Greek symbols
2003.
α convection heat transfer coefficient, kW m−2 K−1 [6] H.M. Jeong, H.S. Chung, E. Djajadiwinata, et al., Experimental study on the
α* dimensionless convection heat transfer coefficient characteristics of longitudinal fin air-heating vaporizers in different seasons,
Γ correction factor J. Mech. Sci. Technol. 22 (5) (2008) 981–990.
[7] H.M. Jeong, Y.H. Lee, M.K. Ji, et al., Natural convection heat transfer estimation
δ fin base thickness, mm from a longitudinally finned vertical pipe using CFD, J. Mech. Sci. Technol. 23
μ dynamic viscosity, Pa s−1 (6) (2009) 1517–1527.
ν kinematic viscosity, m2 s−1 [8] H.M. Jeong, H.S. Chung, S.C. Lee, et al., Optimum design of vaporizer fin with
liquefied natural gas by numerical analysis, J. Mech. Sci. Technol. 20 (4) (2006)
ρ density, kg m−3 545–553.
σ surface tension, N m−1 [9] N. Himoto Advanced design of submerged combustion vaporizer for low
emission operation. The Preliminary Program for 2006 Spring National Meeting.
Japan: LNG III – Environment & Energy. 2006.
Superscript [10] N. Hisada, M. Sekiguchi, Design and analysis of open rack LNG vaporizer,
1 inlet of element ASME/JSME 2004 Pressure Vessels and Piping Conference, Am. Soc. Mech. Eng.
2 outlet of element (2004) 97–104.
[11] T. Jin, M. Wang, K. Tang, Simulation and performance analysis of a heat transfer
tube in SuperORV, Cryogenics 61 (2014) 127–132.
Subscript [12] V.P. Carey, Liquid-vapor Phase-change Phenomena, 2nd ed., Taylor and Francis,
0 outside surface of ice layer 2007.
[13] W. Wilke, Wärmeübergang an rieselfilme, VDI-Forschung. Ver. Dtsch. Ingen.
1 outer tube
490 (4) (1962) 1–36.
2 inner tube [14] R.L. Webb, N.H. Kim, Principles of Enhanced Heat Transfer, Taylor &Francis
g NG Group, New York, 2005.
i internal surface [15] L. Wang, B. Sunden, Performance comparison of some tube inserts, Int. Commun.
Heat Mass Transf. 29 (1) (2002) 45–56.
l LNG or liquid [16] E.W. Lemmon, M.L. Huber, M.O. McLinden, User’s guide, NIST reference fluid
mic microscopic (nucleate boiling) contribution thermodynamic and transport properties – REFPROP, version 9.0, 2010.

View publication stats