Applied Energy 156 (2015) 628–641

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Simulation study on boil-off gas minimization and recovery strategies

at LNG exporting terminals
Yogesh M. Kurle a, Sujing Wang b,⇑, Qiang Xu a,⇑
a
Dan F. Smith Department of Chemical Engineering, Lamar University, Beaumont, TX 77710, USA
b
Department of Computer Science, Lamar University, Beaumont, TX 77710, USA

h i g h l i g h t s

 Various strategies to recover BOG are explored with energy requirement comparison.
 BOG can be recovered to increase revenue of LNG plant and benefit the environment.
 BOG generation can be decreased economically by sub-cooling LNG.
 Heat leaks through LNG equipment are calculated.

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) is becoming one of the prominent clean energy sources with its abundance,
Received 2 April 2015 high calorific value, and low emission and price. Vapors generated from LNG due to heat leak are called
Received in revised form 28 June 2015 boil-off gas (BOG). As world-wide LNG productions are in an increasingly growth, BOG generation and
Accepted 15 July 2015
handling problems become more critical subject to more intense global competitions and stricter envi-
ronmental regulations. In this study, typical C3-MR process, storage facilities, and loading facilities are
modeled and simulated to study BOG generation at LNG exporting terminals, including LNG processing,
Keywords:
storage, and berth loading areas. Factors causing BOG are presented, and quantities of BOG generated due
LNG
Boil off gas
to each factor at each location are calculated under different LNG temperatures. Various strategies to
Flare minimization minimize, recover, and reuse BOG are also studied for their feasibility and energy requirements. The
Natural gas study would help proper handling of BOG problems in terms of minimizing flaring at LNG exporting ter-
C3-MR process minals, and thus reducing waste, saving energy, and protecting surrounding environments.
BOG recovery Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction

With a continuous increase in clean energy demands, the

Abbreviations: BOG, boil-off gas; C3, propane; FBOG, boil-off gas from depres-
surization of LNG after MCHE; FBOG2, boil-off gas from depressurization of
liquefied BOG; FL, BOG generated due to depressurization (flashing) of inlet stream;
HE, BOG generated due to heat added by equipment like pumps; HL, BOG generated
due to heat leak from surrounding into container/pipeline; HT, BOG generated due
to hot tank/container; JBOG, boil-off gas from jetty (while loading a Cargo); LIN,
liquid nitrogen; LNG, liquefied natural gas; MCXB, main cryogenic heat exchanger
bottom section; MCHE, main cryogenic heat exchanger; MCXT, main cryogenic heat
exchanger top section; MR, mixed refrigerant; N2, nitrogen; NG, natural gas; NRU,
nitrogen removal unit used for LNG; NRU2, nitrogen removal unit used for BOG;
TBOG, boil-off gas from LNG storage tanks; VD, BOG generated due to vapor
displacement caused by inlet stream; VRA, vapor return arm.
⇑ Corresponding authors. Tel.: +1 409 880 7798 (S. Wang), +1 409 880 7818
(Q. Xu).
E-mail addresses: swang3@my.lamar.edu (S. Wang), Qiang.xu@lamar.edu
(Q. Xu).
world-wide production capacity of liquefied natural gas (LNG) is
expanding very fast, and LNG is actually becoming the world’s fast-
est growing energy sector. United States Energy Information
Administration (EIA) states that the world natural gas trade, by
both pipeline and shipment in the form of LNG, will be poised to
increase tremendously in the future [1]. 285 million tons per year
(MTPA) of liquefaction capacity has been proposed in North
America alone [2]. New LNG terminals, which are currently under
construction, will increase the LNG production by 125 MTPA [3]. In
2014 only, over 297 MTPA world-wide LNG operating capacity was
recorded [4].
Over long distances, it is more economical to transport natural
gas in the form of LNG, because LNG has over 600 times lower vol-
ume compared with the gas phase of the same mass. However, its

http://dx.doi.org/10.1016/j.apenergy.2015.07.055
0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641
bubble point is below 161 °C, which requires a huge amount of
energy for liquefaction operations. Note that the huge difference
between LNG processing temperature and the ambient tempera-
ture can easily cause heat leak in spite of careful insulations. The
heat leak makes some LNG vaporize, where the vapors generated
are called boil-off gas (BOG). To avoid the overpressure in LNG
containers, it is necessary to relieve BOG periodically. BOG mainly
contains the lightest compounds from LNG, i.e., methane and nitro-
gen. Not having proper BOG recovery facility will lead to flaring of
BOG, which will result in wastage of material and energy, and envi-
ronmental pollutions. Limiting climate change [5] would require
substantial and sustained reductions in greenhouse gas (GHG)
emissions [6]. A range of policies have been made for mitigation
of GHG emissions in different sectors, and these policies are being
implemented effectively by many countries [7]. Various tools for
reduction of GHG emissions include: (a) increase of shares of
renewables, (b) increase of energy efficiency, (c) flare minimization
through proper planning and scheduling operations, avoiding pro-
cess upsets, using better process control, utilizing end flash gases,
(d) minimize venting and fugitive emissions, (e) use of cleaner
fuels, and (f) Carbon capture and sequestration [8]. Methane has
about 26 times higher radiative efficiency than CO2, thus it is more
dangerous to release methane into environment [9]. The global CO2
emissions from flaring of unused gas (natural gas) during oil pro-
duction was about 250 million tonnes in 2011 [10]. Methane was
the second-highest contributor to total GHG emissions during
1990 and 2012 [11]. Therefore, it is important to avoid venting
and flaring of boil-off gas.
LNG industries are actually facing BOG problems in different
sectors of the LNG supply chain: during LNG production, storage,
loading, transportation, and unloading processes. BOG generation
during transportation [12–14] and during unloading [15–18] have
been addressed in many literatures. However, based on the litera-
ture search, it seems that BOG generation at exporting terminals
are still lacking systematic studies. Roughly, BOG generations at
exporting terminals range from 1% to over 3% of the produced
LNG. If they were not recovered and reused, the total amount of
material lost world-wide would be at least equivalent to the capac-
ity of one mid-scale LNG plant. Furthermore, due to more intensive
global competitions and stricter environmental regulations, BOG
flaring is becoming more unacceptable. If this BOG issue at LNG
exporting terminals is not addressed properly and in time, losses
of valuable materials and energy plus air pollutions would be sig-
nificantly greater than ever due to the LNG industry expansions
worldwide. Therefore, BOG minimization at LNG exporting termi-
nals needs special considerations.
In this study, typical C3-MR process, storage facilities, and load-
ing facilities are modeled and simulated to study BOG generation
at LNG exporting terminals, including LNG processing, storage,
and berth loading areas. Factors causing BOG are presented, and
quantities of BOG generated due to each factor at each location
are calculated under different LNG temperatures. Various strate-
gies to minimize, recover, and reuse BOG are also studied for their
feasibility and energy requirements. The study would help proper
handling of BOG problems in terms of minimizing flaring at LNG
exporting terminals, and thus reducing waste, saving energy, and
protecting surrounding environments.
2. BOG minimization and recovery strategies and process
simulation
There are several main LNG processes used in industries: (1)
C3-MR process developed by Air Products & Chemicals Inc.; (2)
Cascade process developed by ConocoPhillips; (3) Duel Mixed
Refrigerant process by Shell; and (4) Mixed Fluid Cascade process
by Linde Engineering. The C3-MR process is used in most LNG
629
plants [19,20]. Therefore, the C3-MR process is used in this study.
Steady-state simulation tool Aspen Plus v8.2 software is used to
simulate NG liquefaction, LNG loading, and BOG recovery
processes.
2.1. Base case simulation
In the base case of C3-MR process, propane is used to precool
natural gas while mixed refrigerant is used for chilling process
for liquefaction. The simulated C3-MR process is partly based on
process flow and process conditions described by Ravavarapu
et al. [21]. Soave–Redlich–Kwong (SRK) cubic equation of state is
used as the property method based on the suggestion of Aspen
Plus for gas processing and hydrocarbon systems. Fig. 1 shows
the process flow diagram for this liquefaction process. An LNG
plant with 4.3 MTPA capacity is simulated using Aspen Plus v8.2.
The feed flow rate is calculated to be 600,000 kg/h. The feed com-
position is given in Table 1. The sweet natural gas enters the plant
at 50 bar and 25 °C. The ambient temperature is set as 15 °C. The
natural gas is precooled to 34 °C using propane refrigeration
cycle. The mixed refrigerant (MR) cycle is also precooled using pro-
pane refrigeration cycle. MR’s model composition includes 40%
methane, 35% ethane, 15% propane, and 10% nitrogen. After the
NG is dried and sent to Scrubber for heavy hydrocarbon removal,
it is directed to nitrogen removal unit (NRU) to separate excess
nitrogen. There are different methods to remove nitrogen from
NG [22–24]. For NRU units in the simulation, it is assumed that
75% of nitrogen from NRU-feed is removed to fuel gas stream com-
ing out of the NRU, and 1.3% of methane from the feed is lost in the
fuel gas stream. After NRU, sweet, dry, and pure NG meeting spec-
ification requirements is sent to bottom section of main cryogenic
heat exchanger (MCHE). It is cooled to 112 °C using heavier part
of mixed refrigerant. At this point, natural gas has been in the liq-
uid form at 49.3 bar. However, this temperature is still higher than
the bubble point of LNG at the storage pressure. Thus, it is further
chilled in top portion of MCHE by the lighter part of MR.
Finally, the LNG exits the main cryogenic heat exchanger at
162 °C and 49 bar with the flow rate of 505,262 kg/h. The compo-
sition of this effluent stream is given in Table 1. The LNG flash, stor-
age, and loading sections are simulated as shown in Fig. 2, where
the LNG stream from MCHE is flashed to depressurize down to
the storage pressure of 1.06 bar at ‘‘DEPRESS’’ tank. The flashing
creates BOG, named as FBOG for the flash tank BOG. The pumping
and piping system is also simulated in order to include hydraulic
calculations, and heat added to LNG by the pumps. The liquid from
the flash tank is then pumped to storage tanks through a 12 in.
pipeline represented by blocks P1 and P2. The equivalent pipe
length from flash tank to storage tanks is assumed to be 1000 m,
with an elevation of 60 m (based on the overall height of the stor-
age tank). The total LNG volumetric flow rate fed to storage tanks is
calculated as 1133 m3/h at 1.06 bar.
Two storage tanks (‘‘STORAGE1’’ and ‘‘STORAGE2’’ in Fig. 2)
with LNG storage capacity 168,000 m3 each are considered in the
study. The settings of the inner diameter of the storage tanks is
70 m; the inner height is 43.75 m, such that the D/H ratio 1.6.
Above ground full-containment type LNG storage tanks is consid-
ered. The tank design, insulation scheme, and heat leak calcula-
tions are explained in Section 2.3.1. The produced LNG is equally
divided and fed to each storage tanks (‘‘STORAGE1’’ and
‘‘STORAGE2’’ in Fig. 2). During the storage, BOG generated from
storage tanks is named as TBOG.
The LNG ship cargo is considered to have four moss type spher-
ical tanks (‘‘SHIP-T1’’ through ‘‘SHIP-T4’’) with a total 143,000 m3
storage capacity. The geometry of the tanks and calculation of heat
leak through these tanks are explained in Section 2.3.2. Note that
each LNG plant can be different in several aspects such as feed
630 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641

NG/LNG
Propane
Mixed Refrigerant
Cooling Water
Air Cooling

Fig. 1. Process flow diagram of LNG plant.

Table 1 originating from the two storage tanks. In Fig. 2, one of these
NG and LNG stream composition.
streams is indicated as ‘‘Note-1’’ to depict its location in the flow
NG (feed) LNG (from MCHE) sheet. In reality, this added amount of LNG is supplied from the
Mass % Mole % Mass % Mole % previously stored LNG. In our simulation, the adjustment is made
that the additional LNG stream would use exactly the same compo-
Methane 80.0 87.48 92.83 96.21
Ethane 6.0 3.50 4.99 2.76 sition, temperature, and pressure as those of the liquid outlet
Propane 2.0 0.80 0.71 0.27 stream of ‘‘STORAGE1’’ and ‘‘STORAGE2’’ in Fig. 2, so that the addi-
n-Butane 1.0 0.30 0.12 0.03 tional stream only adjusts the flow rate of loading line to a fixed
i-Butane 1.0 0.30 0.12 0.03 value by adding the required amount of LNG. This setting is
n-Pentane 0.5 0.12 0.05 0.01
i-Pentane 0.5 0.12 0.05 0.01
dynamic, meaning that change in the properties of liquid outlet
Nitrogen 4.0 2.50 1.13 0.67 stream of storage tank will automatically update properties of
Water 5.0 4.87 0.00 0.00 the corresponding additional stream used to meet a constant
LNG loading rate. This adjustment would not affect the BOG gener-
ation in the storage tank, but will reflect realistic results as long as
and product compositions, methods used for liquefaction, ambient
there is enough LNG flows into the storage tanks to absorb all the
temperature, jetty length, and cargo loading frequency. In this base
heat leak.
case, a long jetty is considered, where the BOG generation is signif-
During the LNG loading, the generated BOG is removed from
icant. Pumps PMP1 through PMP4 are LNG loading pumps. Two
ship tanks to maintain the tank pressure. In real processing, the
pipelines with 24 in. diameter are designated for the LNG loading.
remaining material would accumulate in the ship tank in the form
The equivalent piping length of 6000 m is considered from the
of liquid. Due to limitations of steady state simulation, this liquid
storage area to the ship tanks. The elevation of 50 m is assumed
had to be sent out of the ship tank. As shown in Fig. 2, all four ship
for these pipelines. The piping is represented by pipes P3, P4, P5,
tanks (‘‘SHIP-T1’’ through ‘‘SHIP-T4’’) have provided with liquid
and P6 as shown in Fig. 2. The maximum LNG loading rate is con-
outlet streams in the simulation. This modeling treatment would
sidered to be 10,000 m3/h. This loading rate is used for the simula-
not have any effect on BOG generation from ship tanks but reflect
tion, so that the maximum amount of BOG generation can be
the realistic process. The BOG generated at ship, while loading is in
estimated, and capacities of new BOG recovery equipment can be
process, is named as JBOG for jetty BOG.
identified. It is also assumed that the LNG being loaded is equally
withdrawn from each storage tanks, and each loading line has a
flow rate of 5000 m3/h. 2.2. BOG generation factors
Since the LNG production rate is lower than the LNG loading
rate, the additional LNG stream has to be used to supply the differ- At an LNG plant and exporting terminal, BOG is generated due
ence. Such a stream is used for each of two LNG loading lines to five main factors: (1) depressurization of LNG (flashing); (2) heat
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641
Proposed BOG Recovery Area
Note-1
LNG Plant Area LNG Storage Area
Fig. 2. LNG storage area, loading section, and LNG cargo.
added by equipment like pumps; (3) tank breathing or vapor dis-
placement; (4) heat leaks through containers and pipelines, and
(5) LNG carrying vessels being relatively hot while loading LNG.
The difference between factors (4) and (5) is that for factor (4),
the temperature of the equipment (at a particular point) is almost
constant; whereas in factor (5), equipment needs to be cooled to
achieve isothermal state. For factor (4), heat transfer is from ambi-
ent environment to LNG, and is due to the temperature difference
between LNG/BOG and the ambient. But in factor (5), the heat
transfer is from the container material, and is due to higher tem-
perature of the tank.
In our case studies, feed to each tank (flash tank, storage tanks,
and ship tanks) is set to flash at tank pressure. Pumps and piping sys-
tems are simulated in order to include hydraulic calculations. Thus,
factors (1) and (2) are considered automatically. For factor (3), tank
breathing involves the vapor fed to or removed from a tank to main-
tain its pressure. During the net liquid unloading from a tank, the liq-
uid volume decreased shall be filled with some vapors in order to
maintain the tank pressure, which is called vapor replacement.
Whereas, during the net liquid loading of a tank, the fed liquid will
displace some vapors; these vapors shall be removed from the tank
in order to maintain the tank pressure, which is called vapor dis-
placement. In the simulation, the amount of vapors generated due
to feed flash is already considered in the vapor outlet stream of
the corresponding tank. Thus, to calculate amount of BOG due to
vapor displacement, it requires the addition of some BOG amount
based only on the liquid part of the tank feed. The amount of BOG
generation due to factor (3) is considered in our simulation by using
a calculator block. This BOG amount is directly related to the volume
631
LNG
BOG
Note-2
LNG Vessel
Water Body
LNG Loading Area
of net liquid loading rate. The calculator block is set to add BOG of
volume equal to liquid feed volume at tank pressure (not at feed
pressure). The composition and other properties (e.g., density and
temperature) of the additional stream is copied from the vapor out-
let stream of the corresponding tank. The calculator block is set to
update the values for every change in tank feed condition or any
other simulation parameters. In Fig. 2, one of these BOG addition
streams is indicated by ‘‘Note-2’’ to depict its location. This BOG
stream represents the BOG generation due to vapor displacement.
Such a stream is used for each of the two storage tanks and four ship
tanks. Factor (3) is not applicable to flash tank (‘‘DEPRESS’’ in Fig. 2),
because it does not store LNG, and liquid level in the flash tank is
maintained at a constant value.
In order to cover factor (4) in the simulation, normal designs for
tanks and pipelines are used to calculate overall heat transfer coef-
ficient of the equipment (U). Based on U, surface area, and temper-
ature difference, the total heat leak into equipment can be
calculated in advance and then used as the heat duty for tanks
and pipelines in our simulations. Note that the heat leak through
flash tank (‘‘DEPRESS’’ in Fig. 2) is neglected in this study.
However, the heat leak through LNG storage tanks is not negligible.
The details about heat leak calculations are discussed in the
Section 2.3. Heat leak through pipelines is also calculated based
on overall heat transfer coefficient, size, and temperature gradient.
The heat leak through pipelines is temporarily stored in pipeline
LNG and later results in vapor generation after flashing in storage
tanks or ship tanks. Therefore, vapors generated due to feed flash
in storage and ship tanks represent BOG generation due to pressure
changes, heat added by pumps, and heat leak through pipelines.
632 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641

Factor (5) is applicable to cargo tanks. When cargo tank returns Table 3
from LNG unloading terminal to LNG loading terminal, it carries LNG storage tank – design and materials of construction.

some minimum amount of LNG (called heel) to keep tanks cold. Material of layer (from inside Thermal conductivity Thickness
Still the ship tank temperatures are relatively higher than LNG outwards) (W/(m K)) (cm)
temperature. In this case, the cargo initial temperature of Side wall
100 °C before loading is assumed. The cooling duty required to 1 9% Ni Steel 18.3 4
bring the ship-tanks down to LNG temperature ( 162 °C) is calcu- 2 Insulation – Perlite 0.03 80
3 Pre-stressed Concrete 1.5 50
lated using the following factors – heat capacity and mass of the
tank materials, and the difference between original tank tempera- Roof
1 Aluminum (Flat) 100 0.2
ture (100 °C) and target temperature ( 162 °C). At different time 2 Insulation – Glass Wool 0.05 30
during loading, the heating values would change along with the 3 Insulation – Perlite 0.03 80
LNG loading rate. Note that the maximum LNG loading rate should 4 Reinforced Concrete (ellipsoid 1.5 30
occur only after ship tank cooling is completed to avoid excessive shaped)
thermal shock to the tank material. Bottom
Factor (5), container being hot, is not applicable to flash 1 9% Ni Steel 18.3 0.5
2 Secondary 9% Ni Steel 18.3 0.5
tank or storage tanks onshore. This is because LNG production is
3 Insulation – Cellular Glass 0.045 45
in continuous operation, which makes flash and storage tanks 4 Concrete Base Slab (above 1.5 30
generally work at constant temperatures with respect to time, bottom heater)
and thus require no precooling/cooling, as needed during plant
start-ups. However, the change in temperature gradient due to
listed in Table 3. The average liquid height in the storage tanks is
LNG/BOG/air temperature change will affect heat flux. This
assumed to be 80% of tank inner height. In the simulation, storage
situation has been taken into consideration and are classified
tanks are defined as having flat heads.
under factor (4). The factors considered for all three BOG genera-
LNG, BOG, air, and nitrogen properties used for film coefficient
tion locations are summarized in Table 2. The FBOG amount in
calculations are listed in Table 4. The properties shown for LNG and
Table 2 directly represents BOG only due to depressurization of
BOG are at 162 °C (storage temperature), those for air and nitro-
LNG from 49 bar (production pressure) to 1.06 bar (storage
gen are at 15 °C (the ambient temperature).
pressure).
Note that the LNG loading section is in holding mode when no
2.3.2. Heat leak through cargo tank
ship tank is being loaded; otherwise, it is in loading mode when
LNG cargo with four moss type spherical tanks having the total
ship tanks are being loaded. Based on the production rate and
capacity of 143,000 m3 is considered in this study. Diameter of one
cargo volumes, loading frequency at a loading terminal may vary.
tank is calculated to be 40.9 m. The tank wall is set to be made of
The base case represents the loading mode at peak LNG loading
5 cm thick 9% Ni steel inner layer and 22 cm thick rigid polyur-
rate of 10,000 m3/h at 1.06 bar.
ethane insulation. Thus, the thermal conductivity of 9% Ni steel is
18.3 W/(m K), and that for polyurethane foam is 0.025 W/(m K).
2.3. Heat leak calculation
2.3.3. Heat leak through piping
Heat leak from surroundings into the LNG tank can occur
The average heat transfer coefficients for LNG pipes is assumed
through conduction, convection, and radiation. To calculate con-
to be 0.26 W/(m2 K) [27], and that for BOG pipes is 0.5 W/(m2 K).
vective heat transfer coefficient for various situation, appropriate
Absolute roughness for 9% Ni steel pipe used for LNG flow is taken
empirical correlations [25,26] are used involving dimensionless
as 30 lm, and that for SS/CS pipe used for BOG flow is 45 lm.
numbers: Prandtl number (Pr), Grashof number (Gr), Rayleigh
Details of pipe parameters used are given in Table 5.
number (Ra), Nusselt number (Nu), Reynolds Number (Re).
Iterative calculating method is used to find out layer temperatures.
2.4. BOG minimization strategies
The overall heat transfer coefficients are calculated based on
LNG temperature of 162 °C and air temperature of 15 °C, which
BOG generation is mainly caused by heat leak to LNG in some
is assumed to be constants in all cases of the simulations. The
form. If the LNG temperature is low enough to be able to absorb
change in LNG temperature has negligible effect on U, because
this heat and still remain liquid, BOG generation can be reduced
the insulation resistance has the dominant effect on U compared
significantly. The degree of LNG sub-cooling, and its effect on
to the effect of the change in film coefficients of LNG, BOG and air.
BOG generation have been studied in this section. Sub-cooling
requires more refrigeration energy. Whereas, BOG recycle/recovery
2.3.1. Heat leak through storage tank requires energy for compression and other recovery processes.
Full containment type LNG storage tanks are considered for There should be an optimal temperature of LNG, where the total
heat leak calculations, each with a capacity of about 168,000 m3, cost is at a minimum.
inner diameter 78 m, and inner height of 43.75 m. In general, the
full containment LNG storage tank design includes layers of metal,
2.4.1. LNG sub-cooling in MCHX
insulation, and concrete. The major layers from inside outwards
The effects of sub-cooling LNG on different BOG generation fac-
along with their values of thermal conductivity and thickness are
tors would be different. BOG generated due to flashing can be
reduced by sub-cooling LNG, but that due to vapor displacement
Table 2
Factors considered for each BOG generation location.
is unavoidable. If flashing after MCHE is not intended for nitrogen
rejection, then LNG can be subcooled in MCHE to reduce BOG.
(1) FL (2) HE (3) VD (4) HL (5) HT Power requirement for C3 and MR refrigerant cycles to achieve var-
FBOG Y Y NA N N ious LNG temperatures in MCHX is calculated. For each case of LNG
TBOG Y Y Y Y N temperature from 160 °C to 168 °C, refrigerant quantities are
JBOG Y Y Y Y Y
changed to keep minimum refrigerant flow rate to achieve the
Y – Considered, NA – Not Applicable, N – Not Considered. corresponding LNG temperature.
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641 633

Table 4 pipelines of 24 in. inner diameter and 1000 m equivalent piping

Fluid properties for film coefficient calculations. length, with 20 m elevation from BOG recovery area near the
Parameter Symbol Unit Property values LNG storage area to the LNG plant.
LNG BOG Air N2 TBOG is compressed in CMP-TBOG and sent to the LNG plant
through pipes P11, P12, and P13 of 10 in. inner diameter and
Density q kg/m3 441.5 2.02 1.29 1.2
Viscosity l 10 5 Pa s 14.6 1.1 1.47 1.7
1000 m equivalent piping length with 20 m elevation. When the
Thermal conductivity k W/(m K) 0.22 0.02 0.02535 0.04 BOG recovery area is picked close to the LNG plant, no piping will
Specific heat @ constant Cp kJ/(kg K) 3.6 1.97 1.005 1.04 be considered for FBOG transportation, since the flash tank
pressure (‘‘DEPRESS’’ in Fig. 2) is in the plant area and the required piping
3
Coefficient of volumetric b 10 .1/K 3.78 9 3.5 5.7
length is negligible compared to piping lengths required for
thermal expansion
TBOG and JBOG transfer. FBOG is compressed using compressor
CMP-FBOG. For some other cases, when the BOG recovery area is
picked to be near LNG storage area, FBOG is sent to storage area
2.4.2. LNG sub-cooling after depressurization (final flash) through 1000 m pipeline with 20 m elevation. No piping is needed
In case of flashing BOG after MCHE is necessary for nitrogen for TBOG. JBOG is sent to storage area through VRA, pipe P7, and
rejection, LNG can be subcooled after flash (depressurization). pipe P8 (6000 m piping with 10 m elevation).
However, the single flash does not help much to reduce nitrogen
content. For the base case, LNG has 0.67 mol% after MCHE, which
is reduced to 0.57 mol% after the flash. Sometimes, FBOG might 2.5.1. Use BOG as fuel gas
be desired to meet fuel gas needs of the plant, or for some other LNG plants need a huge amount of energy for compression of
purposes. Huang et al. talk about handling end flash gas and refrigerants and to run other units. BOG can be used as the fuel
designing end flash section of LNG plant to enhance plant effi- to drive compressor and power generation turbines. Fuel gas pres-
ciency and throughput [28]. In case, suppressing FBOG is not sure requirements may vary significantly. Power required to com-
desired, LNG sub-cooling can be performed after the end-flash. press BOG can be calculated.
This sub-cooling will help in decreasing TBOG and JBOG.
2.5.2. Use BOG as feed gas
2.5. BOG recovery strategies BOG can also be used as feed gas for the existing LNG plant. In
this case, the main feed flow rate needs to be reduced to maintain
The following five strategies for BOG recycle/reuse have been the total feed rate within the plant processing capacity. The origi-
considered. nal natural gas feed should be reduced such as to have the total
mass flow rate constant at about 600,000 kg/h. During the LNG
Strategy 1: Use BOG as fuel gas; ship loading mode, the total maximum amount of BOG generation
Strategy 2: Use BOG as feed gas (2a: feed before C3 cycle, 2b: approaches about 10% of the natural gas feed rate by mass accord-
feed before NRU and MR cycle, 2c: Feed BOG to MCHX (after ing to the base case considered. Before the BOG can be used as the
NRU)); feed gas, the BOG stream needs to be compressed to about 50 bar
Strategy 3: Liquefy BOG on shore (3a: using the same refriger- to match the pressure of natural gas stream in the inlet of the
ation cycle, and with NRU; 3b: using the same refrigeration LNG plant. This requires additional energy. The BOG will be com-
cycle, and without NRU; 3c: using separate refrigeration cycle, pressed, air cooled, water cooled, and then recycled to the plant.
and with NRU; 3d: using separate refrigeration cycle, and with- The following three possible locations to feed BOG to the plant
out NRU); are studied.
Strategy 4: Liquefy BOG on berth (4a: using same refrigeration
cycle, 4b: different refrigerant from shore, 4c: refrigerant stored 2a. Feed BOG before propane refrigeration cycle
on berth); and
If BOG is fed to this location, the refrigeration load on C3 and
Strategy 5: Use BOG as make-up gas.
MR cycles should be adjusted properly. This is because although
BOG recovery area can be near storage tank area or near LNG the mass flow rate of the natural gas stream is adjusted to be the
plant area based on the chosen BOG recovery strategy. For same as the base case; however, the total feed composition is chan-
Strategies 1, 2, 3a, and 3b, it can be near plant area. JBOG requires ged. It contained higher amounts of lighter components due to
compression in order to be sent to the BOG recovery area on the BOG recycle. Fig. 3 shows the setup for this strategy.
shore. LNG ships have a compressor already installed aboard with
the lower outlet pressure, e.g., 2 bar. JBOG from ship can be com- 2b. Feed BOG before NRU and MCHX
pressed to 2 bar in CMP-JBOG and sent to the shore through vapor
return arm (‘‘VRA’’ in Fig. 2), and pipes P7 and P8. The pressure In this case, C3 cycle will have no excess load of BOG refrigera-
drop in the VRA is considered to be 0.25 bar for the JBOG flow from tion; meanwhile, the original refrigeration load will decrease if
ship compressor to the base of VRA. Pipes P7 and P8 represent main feed stream flow rate is reduced to keep the total mass flow
pipelines of 24 in. inner diameter and 6000 m equivalent piping in the process stream constant at the normal (base case) flow rate.
length, with 10 m elevation from the base of VRA to BOG recovery MR cycle will experience excess refrigeration load due to the rela-
area close to the LNG storage area. Pipes P9 and P10 represent tively hot (25 °C instead 34 °C) BOG feed, as well as the lighter fed

Table 5
LNG and BOG pipe parameters.

Block name in simulation Location Diameter (in.) Absolute roughness (lm) Pipe material
LNG pipes P1, P2 From plant to LNG storage tanks 12 30 9% Ni Steel
P3, P4, P5, P6 From storage tanks to ship tanks (2 lines) 24 30 9% Ni Steel
BOG Pipes P7, P8, P9, P10 From LNG ship to storage area, and to plant (in series) 24 45 SS or CS
P11, P12, P13 From storage tanks to plant 10 45 SS or CS
634 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641

NG/LNG
C3 Cycle
MR Cycle
BOG Vapor

Fig. 3. BOG as feed gas (BOG recovery strategy 2a).

NG/LNG

BOG

Fig. 4. BOG as feed gas (BOG recovery strategy 2b).

material (methane and nitrogen from BOG). Because of BOG fed to
NRU, the nitrogen recycle to the plant will be less as compared to
not using an NRU. Fig. 4 shows the setup for this BOG recovery
strategy.
2c. Feed BOG to MCHX (after NRU)
Direct BOG feed to MCHE without nitrogen removal is studied
for its feasibility and effects on compression power require-
ments. Precooling of BOG is also tried to reduce the load on
the MR cycle.
2.5.3. Liquefy BOG on shore
After the fulfillment of fuel gas requirements of the plant,
excess BOG can be re-liquefied. This will increase LNG production
of the plant. The same refrigeration cycle used to liquefy natural
gas can be used to re-liquefy BOG, if this increased loads are within
the capacity of the refrigeration units. This strategy saves the cap-
ital investment required to build separate refrigeration facility.
However, it may create controllability issues since the BOG gener-
ation rate changes significantly during LNG loading time periods.
Fig. 5 shows the setup of BOG liquefaction using C3-MR cycles
for Strategy 3a. Fig. 6 shows the use of nitrogen refrigeration cycle
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641
Fig. 5. BOG liquefaction on shore using the same C3-MR cycle (BOG recovery strategy 3a).
[29] for BOG liquefaction for Strategy 3c. For cases 3c and 3d, BOG
recovery area can be close to the LNG storage area, since no refrig-
erant from the plant is being used.
2.5.4. Liquefy BOG on berth
If the BOG liquefaction facility can be installed on the berth
area, jetty BOG can be liquefied and added immediately to the ship
tanks being loaded. This will avoid transportation of BOG back to
the shore. However, ship loadings are in batch operations. Thus,
the corresponding start-up and shut-down of BOG liquefaction fre-
quently might cause problems. Further storage of the refrigerants
will add complexity to the problem. Transfer of on-shore refriger-
ant to berth area has been considered, and energy required for
transportation and heat loss due to additional piping is calculated.
2.5.5. Use BOG as makeup gas
During the loading mode, LNG storage tank needs some gas feed
to maintain its pressure and avoid vacuum operations. The BOG
generation rate in storage tanks is much lower than the volume
of LNG being withdrawn from the tanks during the LNG ship load-
ing. The BOG can be used as make-up gas for the storage tanks.
BOG colder than ambient temperature will add less heat as com-
pared to using other gases at the ambient temperature. The BOG
generation rate is not constant with time, however, JBOG genera-
tion rate is higher than the make-up gas required for storage tanks
as buffers. Therefore, a necessary amount of BOG can be diverted to
storage tanks and the remaining is utilized in other recovery pro-
cesses. This strategy cannot be simulated using steady state
simulation.
3. Simulation results and discussions
3.1. Base case results
For the chosen natural gas composition and the designated flow
rate of 600,000 kg/h, the amount of C3 used is 1,600,000 kg/h, and
635
NG/LNG
C3 Cycle
MR Cycle
BOG Vapor
Liquefied BOG
that of MR is 1,300,000 kg/h. The LNG coming out of MCHX is at
162 °C and 49 bar. The composition of this LNG stream is shown
in Table 1. The LNG leaving flash tank ‘‘DEPRESS’’ contained
methane 92.93%, ethane 5.03%, and nitrogen 0.97% by mass. The
LNG is transferred from the plant to storage tanks by using pump
PMP, which has the outlet pressure of 6 bar. The pressure drop
across pipes P1 and P2 is 4.26 bar. The quantity of LNG fed to
two storage tanks is 1133 m3/h. Net liquid fed to two storage tanks,
i.e., excluding TBOG, is about 1123 m3/h.
LNG at outlet of storage tank contained methane 93.0%,
Ethane 5.08%, and nitrogen 0.84% by mass. Each of pumps used
for LNG ship loading has the flow rate of 2500 m3/h with the out-
let pressure at 10 bar. The total pressure drop across pipes P3
and P4 is about 7.48 bar. Similarly, the total pressure drop across
pipes P5 and P6 is about 7.48 bar. LNG fed to each ship tank is
about 2500 m3/h. Net liquid fed to all the four ship tanks, i.e.,
excluding JBOG, is about 9950 m3/h with the composition of
methane 93.04%, ethane 5.11%, propane 0.73%, and nitrogen
0.76% by mass.
3.2. Quantities of BOG generated
According to the base case simulation, the generated amount of
FBOG is 4897 kg/h with 82.7% methane and 17.2% nitrogen by
mass. TBOG generation is 6905 kg/h with 84.8% methane and
about 15.2% nitrogen by mass. JBOG generation is 45,109 kg/h with
86% methane and 13.9% nitrogen by mass.
Quantity of FBOG depends on LNG composition, temperature,
and pressure. Quantity of TBOG depends on tank feed conditions
as well as several other factors like overall heat transfer coeffi-
cients, surface area of the tank, liquid level in the tank. An 80%
average liquid level is assumed in the steady-state simulation.
Based on design specifications of LNG storage tanks, BOG genera-
tion from the tanks due to heat leaks varies within the range of
0.05–0.1 vol% of the tank per day for different tank designs.
TBOG generation is 6905 kg/h for two storage tanks. Note that
636 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641
BOG Vapor
Liquefied BOG
Nitrogen Cycle
Fig. 6. BOG liquefaction on shore with liquid N2 refrigeration cycle (BOG recovery strategy 3c).
the amount contains vapor displacement by the inflow liquid vol-
ume, which is 1123 m3/h. As the density of TBOG at 1.06 bar is
2.02 kg/m3, the BOG mass due to vapor displacement is calculated
as 2268 kg/h. The feed to the storage tanks is flashed at 1.06 bar.
According to the simulation, it generates 947 kg/h vapors due to
depressurization, heat added by pump, and heat leak through
LNG piping lines. Thus, the BOG from storage tanks due to heat leak
only is about 3690 kg/h, which is obtained by subtracting BOG due
to depressurization and vapor displacement from the total TBOG
amount. This is BOG generated due to heat leak from two storage
tanks. Therefore, the BOG generated in one storage tank equals
1845 kg/h, which is equivalent to 4.18 m3/h (100 m3/day) of
LNG volume based on LNG density of 441.4 kg/m3 at the storage
pressure. For a storage tank with a capacity of 168,000 m3, this is
0.06 vol% of BOG generation per day, which is within the range
of normal tank design specifications.
Jetty BOG amount for the base case simulation is 1.02% of LNG
loaded. Normal range of Jetty BOG generation is reported as 0.6–
Temperature (C)
Pressure (bar)
Mass Flow Rate (kg/hr)
Vapor Fraction
0.8% of transferred LNG per ship loading [30]. Since maximum
instantaneous JBOG generation is considered for the steady state
simulation, it is higher than the reported average value for one ship
loading. Instantaneous jetty-BOG rates vary with LNG loading rate
and ship tank precooling rate. In this study, the maximum JBOG
generation is considered for the study of BOG recovery process,
because design of additional equipment required for BOG recovery
should be based on maximum capacity needs, i.e., the maximum
BOG rates.
3.3. Heat leak calculation results
The calculated heat transfer coefficients for tanks and pipes are
shown in the Table 6. Quantities of heat leaks through pipes are
given from simulation results. For storage tank and ship tank, the
overall heat transfer coefficients are calculated at reference tem-
perature of 162 °C for LNG and BOG. Values for heat leak by radi-
ation are given in Table 7.
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641 637

Table 6
Heat transfer calculations.

Equipment Section UAi (W/m2 K) Ai (m2) DT (°C) QC (kW)

Storage tank Side wall (wetted) 0.049 7697 176.2 66.45
Side wall (non-wetted) 0.048 1924 176.2 16.27
Roof 0.037 3848 176.2 25.09
Bottom 0.066 3848 176.2 44.75
Total – 17,317 176.2 152.56
Ship tank Whole sphere 0.11 5255 176.1 101.79
LNG pipes P1, P2 0.26 958 176.2 43.85
P3, P4 0.26 11,491 175.7 524.81
P5, P6 0.26 11,491 175.7 524.81
BOG pipes P7, P8 0.5 11,491 149 to 119 767.00
P9, P10 0.5 1915 119 to 115 110.77
P11, P12, P13 0.5 798 149 to 134 56.35

UAi – overall heat transfer coefficient based on inside surface area; Ai – inside surface area, DT – temperature gradient; QC – heat transfer by conduction and convection.

Table 7
Heat leaks by radiation (QR) for storage tanks and ship tanks.

Equipment Section Outer surface temperature (°C) Emissivity e QR (kW)

Storage tanks Side wall (wetted) 11.21 0.7 109
Side wall (non-wetted) 11.26 0.7 27
Roof 13.53 0.7 21
Total 157
Ship tanks Whole sphere 7.93 0.8 159

Table 8
LNG sub-cooling effect on BOG quantity and LNG temperature at various locations.

BOG location BOG factor LNG temperature at outlet of MCHE (°C)

160 161 162 163 164 165 166 167
FBOG (kg/h) FL 11,491 8185 4897 1628 0 0 0 0
TBOG (kg/h) FL&HE 943 945 947 948 0 0 0 0
VD 2210 2236 2264 2296 2328 2359 2350 2341
HL 3707 3701 3694 3686 2997 0 0 0
Total 6860 6882 6905 6930 5325 2359 2350 2341
JBOG (kg/h) FL&HE 22,284 22,248 22,210 22,166 22,120 22,080 0 0
VD 19,712 19,816 19,936 20,080 20,236 20,392 20,721 20,720
HL 2974 2969 2963 2955 2950 2945 0 0
Total 44,970 45,033 45,109 45,201 45,306 45,417 20,721 20,720
Total BOG (kg/h) 63,321 60,100 56,911 53,759 50,631 47,776 23,071 23,061
LNG after ‘‘Depress’’ (°C) 161.17 161.27 161.39 161.52 162.09 163.07 164.05 165.03
LNG in storage tank (°C) 161.01 161.10 161.21 161.33 161.46 161.60 162.66 163.63
LNG in ship tank (°C) 160.92 161.01 161.10 161.22 161.34 161.47 161.80 162.78

3.4. Simulation results of BOG minimization strategies
For different LNG temperatures, contribution of each
BOG-generation-factor in generating FBOG, TBOG, and JBOG is
given in Table 8. For different degree of LNG sub-cooling in
MCHE, the corresponding LNG temperatures at various locations
(viz. after flash, in storage tank, and in ship tank) are also given
in Table 8. More sub-cooling of LNG resulted in increased loads
on refrigeration cycles. On the other hand, LNG sub-cooling
decreased BOG, and resulted in less energy requirement for BOG
compression and recovery. There should be an optimum condition
between these opposite trends, where energy requirement is min-
imum. For this, total energy requirement for LNG sub-cooling and
BOG recovery are calculated. The refrigeration power requirement
is based on minimum amounts of refrigerants used to achieve cor-
responding LNG temperature in MCHX. Two separate BOG recovery
processes were chosen – BOG compression to 50 bar to use BOG as
fuel gas (Strategy 1), and BOG liquefaction using Strategy 3a.
Strategy 1 requires lower energy as compared to other Strategies.
Strategy 3a includes BOG compression to 50 bar as well as its liq-
uefaction using C3-MR refrigerants. In Fig. 7, Series-1 (blue line)
shows1 compressor power consumed by refrigerants to liquefy and
sub-cool LNG. This trend goes up with decrease in LNG temperature.
Series-2 (orange trend) shows total power consumption for LNG pro-
cess and BOG compression. Total energy usage increases slightly
with decrease in LNG temperature, up to 165 °C. Then it decreases
drastically till 165.8 °C, and again starts increasing. Series-3 (green
trend) shows total power for LNG process, BOG compression, and
BOG liquefaction using refrigerants. For Series-3, BOG recovery pro-
cess requires relatively high energy, therefore saving this energy by
decreasing BOG amount compensates additional energy spent for
sub-cooling LNG. The trend always goes down with decrease in
LNG temperature, until BOG reaches its lowest possible value.
Therefore, as long as BOG amounts are decreasing, sub-cooling
LNG is always economic in this case. For the both BOG recovery
1
For interpretation of color in Fig. 7, the reader is referred to the web version of
this article.
638 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641

300 Compression Power (kW/ton BOG) Cost ($/ton BOG)

Series1
300 30
Series2 290

Compressor Power (MW)

Compression Power
25

Compression Cost
Series3 250
280
200 20

270 150 15

100 10
260
50 5
250
0 0
0 10 20 30 40 50 60 70
240
-168 -167 -166 -165 -164 -163 -162 -161 -160
Pressure (bar)
LNG Temperature (°C) at outlet of MCHX Fig. 9. Cost of BOG compression.

Fig. 7. Effect of LNG sub-cooling on utility consumption (Series1 – Power required

to compress refrigerants used to liquefy and sub-cool LNG; Series2 – Values from
trends considered, the optimum point for LNG temperature is about
Series1 plus power required to compress total BOG to 50 bar (Strategy 1); Series3 –
Values from Series1 plus power required to compress total BOG to 50 bar and to 165.8 °C.
compress refrigerants used for BOG liquefaction (Strategy 3a)). Fig. 8 shows the change in BOG amounts with change in LNG
temperature, in the form of stacked area chart. FBOG decreases sig-
70,000 nificantly with decrease in LNG temperature and becomes negligi-
Jetty BOG ble at about 163.5 °C. Significant decrease in TBOG starts
60,000
Storage Tank BOG approximately at 163 °C. JBOG starts decreasing approximately
Total BOG (kg/hr)

Flash BOG 50,000 at 165 °C. It can be seen that decrease in total BOG amounts is
significant for LNG sub-cooled between 165 and 165.8 °C. This
40,000
is when the effect of LNG sub-cooling reaches JBOG, and JBOG
30,000 amount decreases. Therefore, the ‘total energy consumption’
trends in Fig. 7 decreases sharply when LNG temperature changes
20,000 from 165 °C to 165.8 °C. Once BOG generation from flashing,
10,000 pumping energy, and heat leaks (Factor (1), (2) and (4)), become
negligible, BOG generation is mainly due to vapor displacement
0 (Factor (3)). TBOG below 165 °C LNG temperature, and JBOG
-168 -167 -166 -165 -164 -163 -162 -161 -160
below 165.8 °C LNG temperature are due to vapor displacement,
LNG Temperature (°C) at outlet of MCHX thus remain almost constant even with decrease in LNG tempera-
Fig. 8. BOG amounts with respect to LNG temperature.
ture. Change in JBOG due to Factor (5), needs to be studied using
dynamic simulation.
Considering time averaged BOG generation rate at exporting
Table 9 terminal in the range of 2–4% of LNG produced, and about 300
Compression duties for FBOG, TBOG, and JBOG. MTPA LNG production worldwide, the range of BOG generation
Flow (kg/h) Duty (kW) Pressure (bar) Temperature (°C) at LNG exporting terminals would be 6–12 MTPA. Fig. 8 and
Inlet Outlet Inlet Outlet
Table 8 show BOG reduction over 60% due to sub-cooling LNG. If
this indicates actual BOG reduction that can be achieved at LNG
CMP-FBOG 4897 75 1.06 2.0 161.4 134.3
CMP-TBOG 6905 107 1.06 2.0 161.2 134.1
exporting terminal, global reduction in BOG can be 3.6–7.2
CMP-JBOG 45,109 703 1.06 2.0 161.1 134.0 MTPA. Methane content of this BOG ranges from 2.7 to 6.9
MTPA. If total BOG at exporting terminals is being flared, the

Table 10
List of additional equipment required for each BOG recovery strategy.

BOG recovery Additional equipment required Energy Required

 100
Energy Recovered
strategy
1 BOG piping, Compressor 1.7a
2a BOG piping, Compressor, Air Cooler, Water Cooler (optional) 5.4
2b BOG piping, Compressor, Air Cooler, Water Cooler (optional) 5.9
3a BOG piping, Compressor, Air Cooler, Water Cooler (optional), NRU, Heat Exchangers – 3, Flash Tank, LNG Pump, LNG piping 5.6
3b BOG piping, Compressor, Air Cooler, Water Cooler (optional), Heat Exchangers – 3, Flash Tank, LNG Pump, LNG piping 5.8
3c BOG piping, Compressor, Air Cooler, Water Cooler (optional), NRU, Heat Exchanger, Flash Tank, LNG Pump, LNG piping 12.8
Nitrogen Refrigeration Cycle (Compressors, Coolers, Expander, JT Valve, Heat Exchangers -2, Storage)
3d BOG piping, Compressor, Air Cooler, Water Cooler (optional), Heat Exchanger, Flash Tank, LNG Pump, LNG piping 13.6
Nitrogen Refrigeration Cycle (Compressors, Coolers, Expander, JT Valve, Heat Exchangers-2, Storage)
4a Refrigerant (Liquid) Piping, Evaporated Refrigerant (vapor) Piping, Flash Tank, Compressor, Cooler, NRU (optional), Heat 6.2b
Exchanger

Notes:
Energy required = Compression cost for BOG and refrigerants for BOG recovery purpose.
Energy recovered = BOG recovered  Methane mass fraction  Calorific value of methane.
a
Calculated for fuel gas pressure of 50 bar.
b
Applicable to JBOG only.
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641 639

Table 11
BOG as feed gas (BOG recovery strategy 2).

Unit Base case BOG recovery strategy

2a 2b
Duty of C3 compressors MW 64.4 64.4 64.8
Duty of MR compressors MW 194.5 205.5 206.7
Duty of BOG compressors MW 13.5 13.7 13.7
LNG in tank (excluding TBOG) kg/h (N2 mass%) 495,726 (0.49) 494,044 (1.09) 494,512 (1.11)
N2 stream from NRU kg/h (N2 mass%) 23,278 (73.5) 29,233 (78.8) 29,815 (79.2)
FBOG kg/h (N2 mass%) 4897 (17.3) 6649 (21.4) 6811 (21.7)
TBOG kg/h (N2 mass%) 6905 (15.2) 6923 (19.0) 6927 (19.3)
JBOG kg/h (N2 mass%) 45,109 (13.9) 45,404 (17.5) 45,429 (17.9)
Total BOG kg/h (N2 mass%) 56,911 (14.4) 58,976 (18.1) 59,167 (18.5)

Table 12
BOG liquefaction on shore (BOG recovery strategy 3).

Liquefied BOG fed to LNG storage tanks? (Y/N) BOG recovery strategy
3a 3b 3c 3d
Refrigerant C3-MR C3-MR LIN LIN
Duty (MW) of refrigerant compressor Base case 258.9 258.9 NA NA
N 298.1 298.4 104.0 104.1
Y 298.0 298.3 103.9 104.0
Duty (MW) of BOG compressors Base casec 13.5 13.5 13.5 13.5
N 13.6 13.5 12.7 12.7
Y 13.6 13.8 12.7 12.7
Power generationa NA NA 10.6 10.6
N2 stream from NRU2 kg/h (N2 mass%) N 6771 (90.6) NA 6771 (90.6) NA
Y 7814 (92.0) NA 7895 (87.0) NA
FBOG2b kg/h (N2 mass%) N 1726 (38.3) 6074 (68.6) 1727 (38.3) 6075 (68.6)
Y 1986 (42.0) 10,759 (82.1) 2007 (42.4) 11,009 (83.0)
TBOG kg/h (N2 mass%) Y 7426 (17.8) 8801 (28.6) 7243 (18.0) 8807 (28.8)
JBOG kg/h (N2 mass%) Y 45,366 (16.6) 46,971 (25.7) 45,333 (16.7) 46,265 (27.0)
Total BOG recycled kg/h (N2 mass%) Y 57,689 (16.6) 60,668 (25.4) 57,473 (16.9) 59,969 (26.5)

NA – Not Applicable.
a
Power generation in MW, from nitrogen expander.
b
BOG generated by flashing of liquefied-BOG only (FBOG from main LNG stream is not included); This FBOG2 has high nitrogen content, thus not recycled to liquefaction
facility.
c
For BOG compression power, all BOG in base case is sent to the plant and compressed to 50 bar to calculate base case BOG compression duty.

reduction in BOG would represent emission reduction of
9–22 Million tonnes CO2 equivalent (Mt CO2-eq) [31] over one year
(assumed 98% flare efficiency), and energy savings of 5–12 GW.
Even if total BOG is being recovered, energy savings in BOG
re-liquefaction due to reduction in BOG amount will be approxi-
mately 0.3 to over 0.7 GW. Thus, sub-cooling LNG can help reduce
emissions as well as costs.
If NG fed to MCHE is already within specification requirements,
particularly in nitrogen amount, FBOG can be theoretically reduced
to zero. If LNG flashing is intended for nitrogen removal, then LNG
can be subcooled after final flash, instead of sub-cooling immedi-
ately after MCHX. End flash changes LNG composition, particularly
in methane and nitrogen content. If assume that LNG with different
composition is flashed to reduce nitrogen content, so that the com-
position after flashing matches that from ‘‘Depress’’ block in Fig. 2,
then Table 8 would represent the TBOG and JBOG quantities cor-
rectly, even for the LNG stream which is sub-cooled after the flash.
3.5. Simulation results of BOG recovery strategies
While sending BOG to LNG liquefaction plant, the pressure drop
for TBOG is 0.13 bar for 10 in. pipeline, and the pressure drop for
JBOG is 0.93 bar for 24 in. pipeline. The compressor duty for each
BOG type is given in Table 9. Each compressor is assumed to have
isentropic efficiency of 0.72 and mechanical efficiency 0.9. The
comparison of energy requirements for various cases (such as –
liquefied BOG recycled or not; NRU used or not) within a
BOG-recovery-strategy, is based on the constant amount of refrig-
erants. The amount of energy required to recover BOG using vari-
ous strategies is given in Table 10. The list of additional
equipment required in each strategy is also given.
3.5.1. Simulation results of BOG usage as fuel gas
Energy utilization and cost associated with BOG compression
from 1.01325 bar (standard atmospheric pressure) to pressures
between 2 bar and 70 bar is shown in Fig. 9. The cost calculation
is based on use of electricity utility at the rate of 0.0775 US $/kW h.
3.5.2. Simulation results of BOG usage as feed gas
The results for Strategies 2a and 2b are given in Table 11. For
BOG recovery strategy 2c, when nitrogen from BOG was not
removed, MCHX could not handle the hot feed, and there would
be temperature cross over occurred at inlet of MCXB. In the case
2c, BOG precooling and nitrogen removal is necessary to avoid
excessive load on MR cycle.
3.5.3. Simulation results of BOG liquefaction on shore
For Strategies 3a and 3b, Fig. 5 shows BOG liquefaction using
the same C3 and MR refrigeration cycles that are used in the base
case. Results are given in Table 12.
640 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641
For Strategies 3c and 3d, Fig. 6 shows BOG liquefaction using
separate refrigeration cycle. Liquid nitrogen is used as the refriger-
ant. The nitrogen expander also drives turbine to generate electric-
ity. The amount of electricity generated is subtracted from energy
required to recover BOG using Strategy 3c and 3d, and net energy
requirement values for these strategies are shown in Table 10. The
results for Strategy 3a through 3d are given in Table 12. For
Strategies 3c and 3d, BOG recovery area is assumed to be close to
LNG storage area. Therefore, JBOG transportation distance
decreases. Thus, the BOG compression cost decreases even when
the BOG amount increases. When BOG is liquefied, it is flashed to
1.06 bar and then sent to storage tanks. The vapors generated after
flashing liquefied BOG contains very high amounts of nitrogen, and
thus it should not be recycled to BOG recovery facility. This vapor
stream needs to be used as fuel gas.
3.5.4. Simulation results of BOG liquefaction on berth
This strategy is similar to Strategy 3, except that the refrigerant
is sent to the berth area instead of transferring BOG to shore. In
Strategy 3, JBOG is transferred to shore through BOG pipeline
and liquefied BOG is again sent to berth for loading using the same
LNG pipelines. If refrigerant is to be pumped to the berth area, it
will require separate cryogenic pipeline, and also the evaporated
refrigerant needs to be compressed and sent back to the plant.
The refrigeration loads would be similar whether BOG on shore
or on berth is liquefied. However, refrigerant needs to be trans-
ported from shore to berth for BOG recovery strategy 4. The capital
cost of cryogenic pipeline might make this strategy costlier. This
strategy is not simulated separately, but the piping for refrigerant
is simulated for sizing purpose. The pressure drop of about 9 bar
with the temperature rise of about 4 °C are observed for 7000 m
equivalent piping length with the elevation of 30 m. The evapo-
rated refrigerant temperature rise is about 50 °C (about 27 °C due
to pipe heat leak, and remaining due to compression), and pressure
drop of 1.25 bar for a 24 in. pipeline.
4. Concluding remarks
This paper has conducted a comprehensive simulation study on
boil-off gas minimization and recovery strategies at LNG exporting
terminals. Factors causing BOG are presented, and quantities of
BOG generated due to each factor at each location are simulated
under different LNG temperatures. Various strategies to minimize,
recover, and reuse BOG are also studied for their feasibility and
energy requirements. BOG generation can be decreased economi-
cally by sub-cooling LNG. The optimum LNG temperature for the
base case in this study was 165.8 °C, which might change with
plant conditions, LNG composition, and BOG recovery method.
Based on energy requirement, use of BOG as fuel gas is the most
efficient way to recover BOG. However, there is constraint for
amount of fuel gas necessary for an LNG plant. Furthermore, this
strategy does not increase plant revenue. Section 3.5.2 explains
use of BOG as feed gas, which saves imported raw natural gas.
Section 3.5.3 considers liquefaction of BOG on shore. This strategy
increases in-tank LNG production, thus generates more revenue.
Section 3.5.4 is about BOG liquefaction on berth, which also
increases plant revenue by increasing LNG export. The studied
strategies show that the energy consumption for BOG recovery is
less than 20% of contained energy of the recovered BOG, which is
equivalent to more than 80% of energy savings when compared
to the complete flaring of BOG. The study shows that BOG recovery
process is economic in terms of operating cost, and has much
scope/potential to be economic in terms of capital cost also. The
study could conceptually help the proper handling of BOG prob-
lems in terms of minimizing flaring at LNG exporting terminals,
and thus reducing waste, saving energy, and protecting surround-
ing environments. It also lays a solid foundation for future dynamic
simulations on BOG minimization and recovery strategies to
explore their controllability and flexibility issues.
Acknowledgements
This work was supported in part by Texas Air Research Center
headquartered at Lamar University and Graduate Student
Scholarship from Lamar University.
References
[1] Barden JL, Ford M. International energy outlook 2013. DOE/EIA-0484(2013);
2013.
[2] Ferrier J. International gas union – world LNG report, 2014 Ed.; 2014.
[3] Conti JJ. Annual energy outlook 2014 with projections to 2040. DOE/EIA-
0383(2014); 2014.
[4] International Gas Union Natural Gas Facts and Figures (World Gas Conference
2015). <http://www.igu.org/sites/default/files/Part%203%28Oct14%29-%
20LNG.pdf> [accessed 16.02.15].
[5] Singh BR. Global warming – impacts and future perspective. InTech: 2012. p
364.
[6] Core Writing Team, Pachauri RK, Meyer LA, editors. Climate change 2014:
synthesis report. In: Contribution of working groups I, II and III to the fifth
assessment report of the intergovernmental panel on climate change. IPCC
2014, AR5.
[7] UNFCCC. Action on mitigation: reducing emissions and enhancing sinks.
<http://unfccc.int/focus/mitigation/items/7171.php> [accessed 25.06.15].
[8] ExxonMobil. Mitigating greenhouse gas emissions in our operations.
<http://corporate.exxonmobil.com/en/current-issues/climate-policy/
emissions-reduction/overview#/section/flaring> [accessed 25.06.15].
[9] Solomon S, Qin D, Manning M, editors. Climate change 2007: the physical
science basis. In: Working groups I contribution to the fourth assessment
report of the intergovernmental panel on climate change. IPCC 2007, AR4.
[10] PBL. Trends in global CO2 emissions: 2013 report. PBL Netherlands
environmental assessment agency, and European commission joint research
center; 2013.
[11] Subsidiary Body for Implementation (UNFCC). National greenhouse gas
inventory data for the period 1990–2012. In: United nations framework
convention on climate change (UNFCC) 2014, FCCC/SBI/2014/20.
[12] Shin Y, Lee YP. Design of a boil-off natural gas reliquefaction control system for
LNG carriers. Appl Energy 2009;86:37–44.
[13] Hasan MMF, Zheng AM, Karimi IA. Minimizing boil-off losses in liquefied
natural gas transportation. Ind Eng Chem Res 2009;48:9571–80.
[14] Romero Gómez J, Romero Gómez M, Lopez Bernal J, Baaliña Insua A. Analysis
and efficiency enhancement of a boil-off gas reliquefaction system with
cascade cycle on board LNG carriers. Energy Convers Manage 2015;94:261–74.
[15] Liu C, Zhang J, Xu Q, Gossage JL. Thermodynamic-analysis-based design and
operation for boil-off gas flare minimization at LNG receiving terminals. Ind
Eng Chem Res 2010;49:7412–20.
[16] Wang M, Xu Q. Optimal design and operation for simultaneous shale gas NGL
recovery and LNG re-gasification under uncertainties. Chem Eng Sci
2014;112:130–42.
[17] Wang M, Zhang J, Xu Q. A novel conceptual design by integrating NGL recovery
and LNG regasification processes for maximum energy savings. AIChE J
2013;59:4673–85.
[18] Jang N, Shin MW, Choi SH, Yoon ES. Dynamic simulation and optimization of
the operation of boil-off gas compressors in a liquefied natural gas gasification
plant. Korean J Chem Eng 2011;28:1166–71.
[19] Wang M, Zhang J, Xu Q. Optimal design and operation of a C3MR refrigeration
system for natural gas liquefaction. Comput Chem Eng 2012;39:84–95.
[20] Wang M, Zhang J, Xu Q, Li K. Thermodynamic-analysis-based energy
consumption minimization for natural gas liquefaction. Ind Eng Chem Res
2011;50:12630–40.
[21] Ravavarapu VN, Oakley JH, White CC. In: Weiss, G., editor. Thermodynamic
analysis of a baseload LNG plant. Chemeca 96: excellence in chemical
engineering; 24th Australian and New Zealand chemical engineering
conference and exhibition; Proceedings. 1996: 143–148 [in volume 1]. no.
96/13. National conference publication: Barton, ACT; Institution of Engineers,
Australia, 1996; Vol. 1, pp 143–143-148.
[22] Kuo JC, Wang KH, Chen C. Pros and cons of different Nitrogen Removal Unit
(NRU) technology. J Nat Gas Sci Eng 2012;7:52–9.
[23] Lokhandwala KA, Pinnau I, He Z, Amo KD, DaCosta AR, Wijmans JG, et al.
Membrane separation of nitrogen from natural gas: a case study from
membrane synthesis to commercial deployment. J Membr Sci
2010;346:270–9.
[24] Rufford TE, Smart S, Watson GCY, Graham BF, Boxall J, Diniz da Costa JC, et al.
The removal of CO2 and N2 from natural gas: a review of conventional and
emerging process technologies. J Petrol Sci Eng 2012;94–95:123–54.
[25] Bengtson HH. Convection heat transfer coefficient estimation. <https://
www.suncam.com/courses/100222-01.html> [accessed 07.01.15].
 Y.M. Kurle et al. / Applied Energy 156 (2015) 628–641 641

[26] Bengtson HH. Natural convection heat transfer coefficient estimation [29] Moon J, Lee YP, Jin Y, Hong E, Chang H. In Cryogenic refrigeration cycle for re-
calculations. <http://www.brighthubengineering.com/hvac/92660-natural- liquefaction of LNG boil-off gas. International cryocooler conference, cryostar.
convection-heat-transfer-coefficient-estimation-calculations/> [accessed the cryostar magazine; 2007. p 629–635.
07.01.15]. [30] Huang S, Hartono J, Shah P. In Recovering boil-off gas during LNG ship loading.
[27] Kitzel B. Choosing the right insulation. <www.phpk.com/pdf/ 86th GPA (Gas Processors Association) Convention. Chevron Company; 2007.
LNGIndustry2008.pdf> [accessed 15.02.15]. [31] Environmental Protection Energy (EPA), United States. Greenhouse gas
[28] Huang S, Tsai N, Oliver S. In simple and versatile methods for modeling LNG- equivalencies calculator. <http://www.epa.gov/cleanenergy/energy-
related systems. In: 88th annual convention of the gas processors association resources/calculator.html#results> [accessed 25.06.15].
2009: 2009 GPA convention – enhancing midstream’s voice. Gas Processors
Association (GPA), vol. 1; 2009.