/smash/get/diva2:426936/FULLTEXT01.pdf

/smash/get/diva2:426936/FULLTEXT01.pdf
Evaluation of Natural Gas Liquefaction
Processes for Floating Applications
Offshore
Øyvind Eckhardt
Master of Science in Product Design and Manufacturing
Submission date: February 2010
Supervisor:
Truls Gundersen, EPT
Norwegian University of Science and Technology
Department of Energy and Process Engineering
Problem Description
The main objective of this Master thesis is to carry out a thorough thermodynamic analysis of a
few suitable liquefaction processes for natural gas in offshore (floating) applications, however,
main emphasis should be placed on the NicheLNG process. The objective also includes
considering ways to improve the energy efficiency of the NicheLNG process.
Assignment given: 16. September 2009
Supervisor: Truls Gundersen, EPT
I
II
III
IV
Preface
This master thesis is a result of the study during my graduation at NTNU. It has been carried out at the
university and at Höegh LNG between September 2009 and February 2010. The focus has been on
thermodynamic qualities and potentials of the NicheLNG process.
I specially want to thank my supervisor at NTNU, Professor Truls Gundersen for his support and
feedback during my work. I would also like thank my industrial contact at Höegh LNG, Vegard
Hellekleiv for given me the opportunity to work on this project. Lastly, I would like to express my
gratitude to Lars Petter Revheim and Thomas Larsen at Höegh LNG for valuable help and
supplementary feedback.
Trondheim, 15.02.2010
____________________
Øyvind Eckhardt
V
VI
Sammendrag
Litteraturstudie av LNG prosesser egnet for offshore har blitt evaluert. Undersøkelsen er utført med
vekt på plassbehov, kompleksitet, effektivitet og sikkerhet. Ved bruk av termodynamikk er NicheLNG
prosessen beskrevet. Flytendegjøringsprosessen på HLNG FPSO-1 har blitt vurdert med hensyn til
energiforbruk, forbedringer og muligheten for å utvide prosessen for å gi en indikasjon på dens
potensial. I tillegg har en alternativ LNG prosess blitt sammenlignet med NicheLNG prosessen.
Prosesser med dobbel ekspander kjølekretser basert på nitrogen som kjølemedium er den mest
foreslåtte løsningen for offshore produksjon. Derfor ble den valgt som en alternativ prosess til
NicheLNG. I vurderingen av de to prosessene ble simuleringer gjort med likeverdige vilkår.
NicheLNG prosessen, basert på en åpen metan krets og en nitrogen krets, hadde en lavere massestrøm
som resulterte i 10% lavere energiforbruk. Valg av kjølemediet (metan eller nitrogen) har forskjellig
spesifikk varmekapasitet og dermed en innflytelse på massestrømmen. For en gitt kjøleytelse krever
metan som kjølemiddel mindre massestrøm enn nitrogen. I tillegg vil høyere trykk nivåer bidra til økt
effektivitet og redusere størrelser på utstyr.
I den åpne kjølekretsen til NicheLNG prosessen blir metan kjølt ned til -1,5°C før den blir ekspandert.
Hvis nedkjølingen blir utvidet til -10°C før ekspansjon er det mulig å oppnå høyere effektivitet for
metan kretsen.
Kriteriene ved sammenligning av LNG prosesser er viktig når kvaliteten skal fastsettes. Fødegassen og
produkt spesifikasjonene gir restriksjoner på oppnåelig effektivitet. Med økende fødegasstrykk kreves
det mindre arbeid (fra fødegass til LNG), men på grunn av den høye virkningsgraden til kompressoren
og den lave virkningsgraden for selve flytendegjøringen vil den totale virkningsgraden reduseres.
Arbeidet flytendegjøringen krever blir mer dominerende enn kompressorarbeidet med høyere
fødegasstrykk. Derfor bør virkningsgraden av LNG prosesser ikke beregnes fra sitt fødegasstrykk men
fra tilstanden etter fødegasskompressor. For NicheLNG prosessen ble eksergi virkningsgraden av
flytendegjøringsdelen beregnet til 26,6%, ved et inngangstrykk på 75 bar.
En vurdering av økt LNG produksjon med vekt på forbedringer for å holde energiforbruket nede ble
undersøkt. De fire undersøkte løsningene var; End Flash Gas, væskefaseturbin, en ekstra kompressor
og økning av varmeveksler areal. Væskefaseturbin var forbedringen som skiller seg ut som den med
høyest bidrag til effektiviteten. Med en 25% økning i LNG produksjon og med de nye enhetene og
modifikasjonene reduserte det spesifikke arbeidet fra 0,5502 kWh / kgLNG til 0,4791 kWh / kgLNG.
Disse forbedringene kan rettferdiggjøre økte investeringskostnader ved 25% høyere LNG produksjon
siden energiforbruket reduserte med 12,9% enn for den opprinnelige utformingen av NicheLNG
prosessen. Grunnet begrenset med plass og vektkapasitet på en FPSO må dette tas hensyn til ved en
forandring av prosessen.
VII
VIII
Summary
A literature survey of LNG processes suitable for offshore environment has been evaluated. The
survey has been performed with emphasis on space requirements, complexity, efficiency and safety.
Thermodynamics theory is described and used in the investigation of the NicheLNG process. The
liquefaction part of the HLNG FPSO-1 has been evaluated with respect to its energy consumption,
improvements and the possibility to expand the process to give an indication about improvement
potentials. In addition, one alternative liquefaction process has been compared with the NicheLNG
process.
Dual expander processes based on nitrogen as refrigerant are the most proposed solution suitable for
offshore applications. Therefore it was chosen as an alternative process to NicheLNG. In the
investigation of the two processes the processes simulated were with equal conditions. The NicheLNG
process, based on an open methane cycle and a nitrogen cycle, had a significantly lower mass flow rate
resulting in 10% lower power consumption. Decision of chosen refrigerant gas (methane or nitrogen)
has different specific heat capacity and hence an influence on the flow rate. Methane as refrigerant
requires less mass flow rate than nitrogen for a given duty. In addition, higher pressure levels will
contribute to increased efficiency and reduced unit sizes.
In the open refrigeration cycle of the NicheLNG process, methane is cooled down to -1,5°C before it
is expanded. If the internal heat exchange is extended to -10°C before expansion is it possible to
achieve some efficiency increase for the methane cycle.
Comparison criteria are important when the quality of liquefaction processes is to be determined. Feed
and product specifications provide some restrictions on obtainable efficiency. With an increasing feed
gas pressure, the whole liquefaction process (from feed to LNG) demands less work, but the overall
process efficiency is reduced due to the high efficiency of the feed gas compressor and the low
efficiency of the liquefaction part. Hence, the efficiency of the liquefaction process should not be
calculated from its feed gas pressure but rather the liquefaction pressure. For the NicheLNG process,
the exergy efficiency of the liquefaction part was calculated to 26,6%, with a liquefaction pressure at
75 bar.
An increase of LNG production with emphasis on improvements to keep work consumption down was
also discussed. The four evaluated solutions were utilization of End Flash Gas, liquid expander,
additional compressor and increase of heat exchanger area. The liquid expander was the improvement
that stands out as the highest contribution to the efficiency. With a 25% increase in LNG production
and with new units and modifications of the design resulted in a reduction in the specific work
consumption from 0,5502 kWh/kgLNG to 0,4791 kWh/kgLNG. These efficiency improvements can
justify higher investment costs since the work consumption, with 25% higher LNG production, was
12,9% lower than for the original design of the NicheLNG process. Never the less, space and weight
on a FPSO are limited and has to be considered when a more efficient process is desired.
IX
X
Preface ................................................................................................................................................... V
Sammendrag ...................................................................................................................................... VII
Summary .............................................................................................................................................. IX
List of figures .................................................................................................................................... XIII
List of tables ........................................................................................................................................XV
Nomenclature .................................................................................................................................... XVI
Abbreviations ................................................................................................................................. XVI
Letters ............................................................................................................................................ XVII
Prefixes .......................................................................................................................................... XVII
Introduction ...................................................................................................................................... XIX
1
Theory............................................................................................................................................. 1
1.1
Compression and expansion .................................................................................................... 1
1.2
Coefficient of Performance ..................................................................................................... 2
1.3
Brayton Refrigeration Cycle.................................................................................................... 4
1.3
Exergy ..................................................................................................................................... 4
1.4
Heat Exchanger Duty .............................................................................................................. 7
1.5
Refrigerant medium ................................................................................................................. 8
1.5.1
Gaseous refrigerants ........................................................................................................ 8
1.5.2
Mixed refrigerants ........................................................................................................... 9
1.6
2
3
Simulation specifications....................................................................................................... 10
Description of liquefaction technologies .................................................................................... 11
2.1
Niche LNG ............................................................................................................................ 13
2.2
PRICO - Single Mixed Refrigerant ....................................................................................... 14
2.2.1
Principle......................................................................................................................... 14
2.2.2
Extensions of PRICO .................................................................................................... 14
2.3
Dual nitrogen refrigerant ....................................................................................................... 15
2.4
Comparing conventional with expander liquefaction processes ........................................... 16
Promising liquefaction processes for FPSO applications ........................................................ 17
3.1
The NicheLNG process ......................................................................................................... 17
3.1.1
3.2
Dual nitrogen process ............................................................................................................ 22
3.2.1
3.3
4
Exergy analysis of the NicheLNG process .................................................................... 19
Exergy analysis of a dual nitrogen process ................................................................... 22
Discussion of NicheLNG versus dual N2 process ................................................................. 24
Adjustments and analysis of NicheLNG.................................................................................... 25
4.1
Precooling.............................................................................................................................. 25
XI
4.2
4.2.1
Refrigerant medium ............................................................................................................... 26
Change in gas characteristics for pressure variations .................................................... 27
4.3 Placement of expansion ............................................................................................................... 29
4.3.1
Expander placement for the open methane cycle .......................................................... 29
4.3.2
Consequences of expander placement ........................................................................... 30
4.4
4.4.1
4.5
5
Liquefaction pressure ............................................................................................................ 31
Relationship between feed gas and liquefaction pressure ............................................. 31
Discussion on the analysis ..................................................................................................... 35
Increased capacity of the NicheLNG process............................................................................ 37
5.1
Utilization of end flash gas (EFG)......................................................................................... 37
5.2
Liquid expander ..................................................................................................................... 38
5.3
Two stage compression ......................................................................................................... 40
5.4
The improvements influence by higher LNG production ..................................................... 41
5.5
Discussion on increased capacity .......................................................................................... 45
6
Conclusions and further work.................................................................................................... 47
6.1
Conclusions ........................................................................................................................... 47
6.2
Suggestions on further work.................................................................................................. 48
REFERENCES .................................................................................................................................... 49
Appendix A .......................................................................................................................................... 50
Appendix B........................................................................................................................................... 54
Appendix C .......................................................................................................................................... 55
Appendix D .......................................................................................................................................... 56
Appendix E-1 ....................................................................................................................................... 57
Appendix E-2 ....................................................................................................................................... 58
XII
List of figures
Figure 1.1
Enthalpy – entropy diagram. Isenthalpic and isentropic expansion illustrated with
arrows [12],
page 1
Figure 1.2
A refrigeration process [13], page 2
Figure 1.3
Coefficient of Performance as a function of temperature, page 3
Figure 1.4
Carnot reversed cycle, page 3
Figure 1.5
Flow sheet and T-s diagram of a reversed Brayton cycle [18], page 4
Figure 1.6
Simple illustration of an expander refrigeration process, page 5
Figure 1.7
Counter current heat exchanger, page 7
Figure 1.8
Temperature-enthalpy diagram of natural gas with cooling sequences indicated, page 8
Figure 2.1
Flow sheet of the NicheLNG process [5], page 13
Figure 2.2
Basic principle of a single mixed refrigerant process [7], page 14
Figure 2.3a
Statoil proposed solution, page 15
Figure 2.3b
BHP Billiton proposed solution, page 15
Figure 3.1
Pressure – temperature diagram illustrating the natural gas path [5], page 17
Figure 3.2
Pressure – temperature diagram illustrating the nitrogen path [5], page 18
Figure 3.3
Temperature profile in the heat exchanger, page 18
Figure 3.4
Specific work as function of overall exergy efficiency for a LNG process from
gaseous feed at different pressures to saturated liquid at 1 bar, page 19
Figure 3.5
The composite curves for the NicheLNG process, page 21
Figure 3.6
Flow sheet of the simulated NicheLNG process, page 21
Figure 3.7
The composite curves for the dual nitrogen process, page 23
Figure 3.8
Flow sheet of the simulated dual nitrogen process, page 23
Figure 4.1
The NicheLNG process with a precooler in front [17], page 25
Figure 4.2
Simple illustration of expander precooling, page 26
Figure 4.3
Specific heat capacities of N2 and CH4 at pressure levels from the NicheLNG process,
page 27
Figure 4.4
Specific volume variations with pressure at a temperature of 30 °C, page 28
Figure 4.5
The NicheLNG open methane refrigeration cycle with refrigeration regions indicated,
page 29
XIII
Figure 4.6
Efficiency of the open natural gas cycle with different inlet temperatures, page 30
Figure 4.7
Natural gas path through liquefaction for a typically onshore facility [11], page 31
Figure 4.8
Ideal liquefaction process of natural gas [11], page 32
Figure 4.9
Simple flow sheet of the liquefaction path from feed gas to LNG, page 32
Figure 4.10
Specific work of compression from 1 bar to a certain liquefaction pressure and specific
liquefaction work, page 33
Figure 4.11
Work consumption for compression from a feed gas pressure to a liquefaction pressure
and a 26,6% efficient liquefier, page 34
Figure 5.1
Pressure-temperature diagram for pressure reduction with valve and liquid expander
followed by a valve, page 38
Figure 5.2
A graphical overview work saved by integration of a liquid expander, page 39
Figure 5.3
25% higher LNG production with increase of UA-value, page 41
Figure 5.4
The extended NicheLNG process, page 42
Figure 5.5
Temperature-enthalpy diagram of the 2DLE with 25 % increased capacity, page 44
Figure 5.6
Temperature-enthalpy diagram of the Extended 2DLE with constant LMTD and 25 %
increased capacity, page 44
XIV
List of tables
Table 1.1
Exergy losses in different components [3], page 6
Table 1.2
Natural gas mole% composition, page 10
Table 2.1
Liquefaction processes suitability for FPSO [9], page 11
Table 2.2
Efficiency table of expander natural gas liquefiers [4], page 12
Table 3.1
Efficiency and work consumption for the NicheLNG process, page 20
Table 3.2
Minimum liquefaction work to feed gas pressure, page 20
Table 3.3
Efficiency and work consumption for the dual nitrogen process, page 22
Table 3.4
Results of both processes, page 24
Table 4.1
Specifications and results of a precooling example, page 26
Table 4.2
Results of inlet temperature as original at -1,5°C and at -10°C, page 30
Table 4.3
Exergy calculations of the liquefaction part (without feed gas compressor and for a
single train), page 33
Table 5.1
COP of 2DLE and the EFG, page 37
Table 5.2
Results from the integration of a liquid expander, page 40
Table 5.3
Individual improvements in efficiency with new a unit or change in the design (EFG),
page 42
Table 5.4
Results of 25 % increase in LNG production, page 43
XV
Nomenclature
Abbreviations
BTU
British Thermal Unit
CB&I
Chicago Bridge & Iron
CH4/C1
Methane
C2H6/C2
Ethane
C3H8/C3
Propane
CO2
Carbon dioxide
Comp.
Compressor
COP
Coefficient of Performance
C3MR
Propane Mixed Refrigerant
DMR
Dual Mixed Refrigerant
EFG
End Flash Gas
FEED
Front End Engineering Design
FPSO
Floating, Production, Storage and Offloading vessel
FPSO-1
The first planned FPSO
HC
Hydrocarbon
HLNG
Höegh LNG
HX
Heat Exchanger
LMTD
Log Mean Temperature Difference
LNG
Liquefied Natural Gas
LPG
Liquefied Petroleum Gas
NG
Natural Gas
N2
Nitrogen
N2-Exp
Nitrogen Expander
O2
Oxygen
Scf
Standard Cubic Foot
SMR
Single Mixed Refrigerant
SRV
Shuttle and Regasification Vessel
T-s diagram
Temperature-entropy diagram
2DLE
NicheLNG original design
XVI
Letters
A
Area
Cp
Specific Heat Capacity
e
Specific Exergy
F
Correction factor
h
Enthalpy
m
Mass rate
n
Mole rate
P
Pressure
s
Entropy
T
Temperature
T0
Ambient temperature
Q
Heat duty/rate
U
Heat Transfer Coefficient
W
Work
µJ
Joule-Thomson Coefficient
η
Efficiency
Prefixes
103
k
kilo
M
Mega 106
G
Giga
109
XVII
XVIII
Introduction
For many years, several companies have been working on realization of offshore LNG production. A
proposed production facility is an LNG Floating, Production, Storage and Offloading vessel (FPSO) as
foundation for the liquefaction application. Destinations of an FPSO are isolated gas field. In gas
fields remote from land may it be uneconomic to build up an infrastructure to exploit the reserves. In
addition is associated gas from offshore oil production that is flared or re-injected into reservoir a
possible placement.
At the present time, Höegh LNG is operating LNG ships for transport and is about to hand over two
SRV ships (Shuttle and Regasification Vessel). The company wants to have a solution in a floating
value chain for LNG with integration of a LNG production facility. The development of an FPSO for
LNG production has reached the end of the FEED (Front End Engineering Design) phase. The design
is not yet set since final destination is unknown and an increase in production rate may be desired. A
possibility of higher LNG production or better energy efficiency can give an advantage in the future
customer negotiations.
This master thesis takes the original design of the NicheLNG process and compares it with other
liquefaction processes suitable for offshore applications. Restricted to offshore production, some
aspects have higher importance than for land based facilities. Potentials and improvements are
analyzed both from a practical viewpoint but also some hypothetical situations are discussed. Since
detailed information on the processes is restricted, some considerations have been made to get equal
assumptions to compare processes. This will be expressed later in the thesis.
The theory of the underlying thermodynamic calculations is expressed followed by a literature survey
of liquefaction processes suitable in an offshore environment. These processes are then compared in
energy efficiency, chosen design and refrigerant medium. Then the improvement potential with
emphasis on the most promising process is analyzed. An indication of the potential expansion in
capacity is done by manipulating refrigerants and with a higher equipment count. In contrast to the
original design analysis, the evaluation of increased capacity has a more practical view.
Two energy saving sources have in an earlier master thesis been identified and evaluated, and will
therefore not be treated in this work. It covered the benefit in terms of energy consumption with lower
cooling water temperature and removal of the NGL extraction process.
XIX
XX
1
Theory
Calculations performed in this thesis are based on well known thermodynamics and simulations in
AspenTech HYSYS. This chapter provides an overview of the thermodynamic principles with
emphasis on cryogenic processes.
1.1
Compression and expansion
Compressors used in the liquefaction section are the main consumer of energy. The amount of
consumed work depends on the inlet and outlet state. From the first law of thermodynamics [1]:
i
i


i
i
i 
i 
dEcv
V
V
i
e




= Q- W + ∑ mi hi +
+ gzi - ∑ me he + + gze



 e
dt
2
2
i




(1.1)
For a steady state system with one inlet, one exit and neglecting change in potential energy, the energy
balance is reduced to:
i

V
i
0 = Q- W + m  hi +

2

i
i
i
i

 i
V
 - m  he + e 
 
2
 

(1.2)
When accounting for entropy, enthalpy and pressure changes, work can be expressed by the magnitude
of the specific volume of the fluid (∆ke = ∆pe = 0) [1]:
 i
 Wcv
 mi


e

= -∫ vdp

i
int rev
(1.3)
Figure 1.1 Enthalpy – entropy diagram. Isenthalpic and isentropic expansion illustrated with arrows [12]
Figure 1.1 illustrates throttling and expansion of a gas. From the relation between the enthalpy change
for real and isentropic expansion, ∆he/∆hs, the efficiency can be found. In cryogenics the
1
efficiency is defined as the reduction of the enthalpy and not the gained mechanical power [12]. An
expansion without an enthalpy reduction is achieved by throttling. The expansion absorbs no heat and
does no work. An isenthalpic expansion is defined by the Joule-Thomson coefficient:
 ∂T 
µJ =  
 ∂p h
(1.4)
A gas that is cooled through a valve has a positive µ J. When negative µ J occur, the temperature
increases by an expansion. The phenomenon of increased temperature is of particular importance
when handling gases at very low temperatures such as helium.
An isentropic expansion is also illustrated in figure 1.1. The isentropic expansion corresponds to a
process with no internal irreversibilities. This is an ideal expansion where maximum achievable work
is developed. In a real expander, entropy is produced resulting in a higher outlet enthalpy value. Thus,
less work is produced.
1.2
Coefficient of Performance
A refrigeration process withdraws heat at rate QC
from a cold source at temperature TC. This is then
delivered to a warmer reservoir at rate QH and
temperature TH. To accomplish this, work input is
necessary. As figure 1.2 illustrates, required work to
perform this cooling is W=QH-QC. Higher TC results
in higher QC, and thus lower work consumption.
With increasing TC the efficiency of a refrigeration
process is increased, as eq. 1.5 and 1.6 express.
Figure 1.2 A refrigeration process [13]
The efficiency of a refrigeration process is commonly defined by a Coefficient of Performance (COP),
and is defined [1]:
COP =
QC
W
(1.5)
and theoretically maximum [1]:
COPth,max =
2
1
 TH 
  -1
 TC 
(1.6)
From Eq. 1.6 and with hot reservoir temperature at 38°C as TH=T0, the maximum COP as a function of
cold temperature TC is plotted in Figure 1.3. The cold temperature TC is the temperature the cooling
duty is delivered.
Figure 1.3 Coefficient of Performance as a function of temperature
The dotted line in figure 1.3 expresses the COPth,max at TH=15 °C. A lower sea water temperature
results in a higher COP, hence a more efficient refrigeration process.
In refrigeration processes, the reversed Carnot
cycle can be used as an illustration of the
theory. An ideal gas with heat absorption and
rejection at constant temperatures is illustrated
in the T-s diagram in Figure 1.4. The reversible
process in Figure 1.4 is stated through
isentropic compression (1-2), isothermal
compression (2-3), isentropic expansion (3-4)
and isothermal expansion (4-1). Work can then
be represented as W and heat extracted as QC in
Figure 1.4.
The COP for the Carnot cycle is expressed as in
Eq. 1.6.
Figure 1.4 Carnot reversed cycle
3
1.3
Brayton Refrigeration Cycle
A Brayton refrigeration cycle is a reversed Brayton cycle. Heat is transported from a cold reservoir,
where the temperature after an expansion is below that of the cold reservoir. Refrigeration is then
achieved by attracting heat from a cold region and then later released. Different from other
refrigeration cycles this cycle involves no phase change. The working fluid remains as a gas
throughout.
Figure 1.5 Flow sheet and T-s diagram of a reversed Brayton cycle [18]
Figure 1.5 illustrates a refrigeration process based on the reversed Brayton cycle. The T-s diagram
represents a real cycle. An ideal cycle operates with isentropic turbine and compressor. As can be
seen, the expansion and compression is not isentropic so a real process operates with some losses.
Heat transfer from the cold region is from 5-6 and then the heat is released after a compression from 23.
1.3
Exergy
Exergy is a measure of the maximum amount of work that can be extracted from a process stream
when it is brought to equilibrium with is surroundings in a hypothetical reversible process [1]. When
neglecting changes in composition and chemical exergy, this is a measure of the potential in thermomechanical exergy and thus defined only in terms of the stream enthalpy and entropy relative to the
surroundings. The exergy, e, expressed at steady-state conditions and neglecting kinetic and potential
energy [2]:
e = ( h - T0 s ) T , P -( h - T0 s )T0 P0
(1.7)
where T0 and P0 are at ambient conditions. When taken from one state to another, the change in exergy
is given by:
∆e = ( h - T0 s ) T2 , P2 -(h - T0 s )T1P1
4
(1.8)
In a real process irreversibilities exist. So actual work required to bring a process to a state is more
than in an ideal case. Given by the second law of thermodynamics, over an actual system, lost work
from compression can be defined as the difference between actual work required and the change in
exergy:
Wlost = Wactual - ∆e
(1.9)
and lost work from expansion:
Wlost = ∆e − Wactual
(1.10)
When exergy production and losses are known the exergy efficiency can be decided. Exergy efficiency
is defined as the relation between the exergy change of natural gas to be liquefied and the power
consumed. Exergy efficiency is defined as:
ηex =
Minimum power for liquefaction
Power Consumption
(1.11)
Figure 1.6 Simple illustration of an expander refrigeration process
Figure 1.6 shows a simple a refrigeration process, consisting of a compressor, aftercooler, heat
exchanger and expander. The compressor consumes work by increasing the pressure of a refrigerant to
a desired level. Necessary cooling is obtained by expanding the refrigerant through a turbine and this
will also produce work. From consumed and produced work the exergy efficiency can be expressed as:
i
ηex =
n1 (e2 - e1 )
i
i
(1.12)
-W comp - W exp
According to [3], the placement of state 2 is in front of the valve when exergy change from feed to
after heat exchanger is to be calculated.
5
Calculations of exergy losses through the components are done as expressed in Table 1.1:
Table 1.1 Exergy losses in different components [3]
From Table 1.1 the exergy losses of components in a liquefaction system can be found, and thereby
the exergy efficiency of a component. Comparing different liquefaction processes the exergy losses of
components in each process may differ and a more detailed overview of the processes may give
advantages in optimizing them. With an overview of the exergy losses from each component may an
optimization be easier to accomplish.
6
1.4
Heat Exchanger Duty
Transfer of heat through a heat exchanger is typically done for three different reasons. Either a stream
needs to be heated or cooled, or a liquid stream needs to be vaporized, or a vapor stream needs to be
condensed. To transfer heat, one rule has to be satisfied from the Second Law of Thermodynamics;
heat can only be transferred from a higher temperature to a lower one. This means that the higher
temperature cooling curve and the lower temperature heating curve cannot intersect.
Figure 1.7 Counter current heat exchanger
The duty of a heat exchanger as in Figure 1.7 can be provided from an energy balance. If inlet and
outlet conditions of a hot stream are known and the specific heat capacity assumed to be constant, the
duty can be expressed as [6]:
i
i
Q H = m Cp H ( TH ,in - TH ,out )
(1.13)
or similarly for a cold stream:
i
i
Q C = m CpC (TC ,out - TC ,in )
(1.14)
To determine the temperature driving force for heat transfer, the log mean temperature difference,
LMTD, is used. The use of LMTD is valid both for co-current and counter-current flow [6] as long as
the specific heat capacities and the heat transfer coefficient are constant. By defining the temperature
difference for each side of the heat exchanger, the LMTD is defined as follows [6]:
LMTD =
∆TL - ∆TR
 ∆T 
ln  L 
 ∆TR 
(1.15)
where ∆TL is the temperature difference on the left side and ∆TR on the right side of the heat
exchanger.
When looking at a cross-flow heat exchanger a common principle is to introduce a correction factor, F.
The correction factor is in the 0 < F ≤ 1 region.
With conduction and convection coefficients, the overall heat transfer coefficient, U, can be
introduced. For a given heat transfer area, A, the duty of a heat exchanger can then be expressed as
[6]:
i
Q = U × A × LMTD × F
(1.16)
7
1.5
Refrigerant medium
A traditional large onshore liquefaction plant normally contains three refrigeration cycles. Each cycle
is supposed to cover different temperature regions of the natural gas to be liquefied. The purpose of
the cycles is precooling, liquefaction and subcooling. When it is desirable to have two cycles the first
operates as a precooler followed by a second cycle for the liquefaction and subcooling.
1.5.1 Gaseous refrigerants
If a gas is chosen as refrigerant without any phase change, operation in the cold end differs often from
a mixed refrigerant. A gaseous refrigerant has the possibility of expanding through a turbine and
thereby produce some work in addition to lower its temperature. Placement of expansion is complex
and depends on the overall process and the chosen refrigerant gas. Decision of refrigerant gas and
placement of expander are covered in Chapter 4.
Figure 1.8 Temperature-enthalpy diagram of natural gas with cooling sequences indicated
Figure 1.8 shows the enthalpy-temperature variation for natural gas at a pressure of 19,5 bar and 75
bar. The composition of the natural gas is a typical feed gas before a liquefaction process. The isobar
lines are chosen from the NicheLNG operating pressures. The 75 isobar line is the liquefaction
pressure and the high pressure level of the methane cycle.
The main objective is to obtain composite enthalpy-temperature variation of hot and cold streams as
close to parallel as possible. The 75 bar slope shows that the specific heat, cp or (∆h/∆T)p, differ with
reducing temperature. As can be seen, the natural gas at 75 bar has three different gradients. Hence,
utilization of three refrigerant cycles with different composition will be the best option.
Figure 1.8 also shows that the feed gas pressure influence the slope with reducing temperature. The
chosen pressure depends on the composition of the feed gas. Higher fraction of heavier hydrocarbons
results in higher feed gas pressure to avoid early entrance in the two phase area. Even though feed gas
8
may differ with different liquefaction processes the variation cannot be large since the end product
specifications (composition and higher heating value) have to be almost equal.
The discussion above is done with pretreatment in mind before natural gas is fed into the liquefier.
Some liquefaction plants have the removal of impurities, as CO2 and water, and/or heavy
hydrocarbons integrated in the liquefaction process. Integration will influence the chosen feed
pressure.
1.5.2 Mixed refrigerants
A liquid containing a pure refrigerant, as propane, evaporates at constant temperature. With single
component evaporation the temperature profile will be horizontal as long as some liquid remains. If a
fluid has a mixed composition is the temperature profile depending on the chosen fluids and the
mixture composition. The mixed refrigerant must contain fluids with boiling temperatures which cover
the whole temperature range. Natural gas to be liquefied needs to be cooled from an ambient
temperature to -162°C. To cover this range, a composition of different hydrocarbons and nitrogen is
often used. Concepts of different mixed refrigerants processes are either as a single mixed cycle or as a
mixed cascade cycle. A mixed refrigerant enables the temperature profile of the cold and hot streams
to be as close as possible. Closer temperature profile results in reduced compressor power and higher
exergy efficiency.
9
1.6
Simulation specifications
Except for some energy balances and exergy calculations most of the results in this thesis are based on
simulations in Aspentech HYSYS. The data for the NicheLNG process was provided by Höegh LNG
and included in the HYSYS file 2DLE. This file is a model of the NicheLNG process after the LPG
fractionation, where lean natural gas is cooled and ends as LNG. The 2DLE file is defined as the
original case and will be referred to as 2DLE.
Efficiencies, ambient conditions, outlet aftercooler temperature
and gas compositions are in 2DLE defined and used as basis in the
other investigated cases. The compressor efficiencies are defined
with vendor curves and the expanders have an adiabatic of 87%.
In order to simplify the analysis was polytropic efficiency at 82%
chosen instead of vendor curves. Cooling water is able to cool
down the streams to 38°C and the streams undergo a pressure drop
of 30 kPa through the aftercoolers. The feed gas to be liquefied
has the composition expressed in Table 1.2. As Table 1.2 shows,
the nitrogen content is close to 2,3 mole%. The desired content of Table 1.2 Natural gas mole%
LNG is below 1 mole% and a higher heating value <11,074 composition
kWh/m3 (<1070 BTU/scf). Due to the volatility of nitrogen its
content can be reduced by production of flash gas. To be able to meet the LNG specifications, the
natural gas leaving the cold box is cooled down to the same temperature in all investigated cases. One
exception is when a liquid expander is introduced in Chapter 5. The nitrogen refrigeration cycles have
a composition of 98 mole% nitrogen and 2 mole% oxygen. The FPSO-1 is supposed to have two
identical liquefaction trains. In this thesis, the evaluations and analysis will be on a single train.
10
2 Description of liquefaction technologies
Natural gas from a reservoir may have to undergo cleaning and scrubbing before liquefaction. This is
necessary if the natural gas contains impurities and do not satisfy the product specifications of LNG.
Heavy hydrocarbons, nitrogen, mercury, water and energy content are important specifications for the
liquefaction process and the customer. Once the natural gas specification has fulfilled the requirements
it is ready for the final liquefaction stage. The liquefaction process is based on the gas being cooled to
its condensation temperature -162°C at atmospheric pressure. This temperature is defined to lie within
cryogenic temperatures. By converting natural gas to liquid state, the gas volume is reduced to almost
1/600 [14]. This enables efficient storage and transportation.
Liquefaction processes in operation have a wide range of complexity. They differ in efficiency and
size. By adding units or cycles, the efficiency may increase, but the size and weight will also increase.
Onshore facilities do not have the strict constraints of low weight and small size so their efficiency is
higher and the production rate too. They can have higher equipment count and an opportunity of large
amount of hydrocarbon storage. These constraints are of importance when designing an offshore
facility. Space is limited and the use of hydrocarbons should be inherently safe. It must also offer a
high degree of modularity, low equipment count, quick start-up, available and be robust to vessel
motion.
Liquefaction processes are either based on cascade, mixed or pure refrigerants cycles. The number of
cycles differs from one to three and is of importance in the success of an efficient liquefaction.
Proposed processes for offshore applications often involve one or two cycles.
Table 2.1 Liquefaction processes suitability for FPSO [9]
Table 2.1 illustrates the most important selection criteria for a liquefaction process for natural gas.
This is only a rough indication of the challenges each process face. This thesis will not cover cascade
cycles and mixed refrigerant cycles with three stages, since these most likely are not suitable for
offshore LNG production. Table 2.1 indicates expander processes (N2-Exp and NicheLNG) as the
most suitable ones for an offshore environment.
Several liquefaction processes have been proposed for an FPSO. These processes range from one to
two mixed refrigerant cycles or expansion cycles involving pure refrigerants. A typical single mixed
refrigerant (SMR) process is the well known PRICO process from Black & Veatch. This refrigerant
cycle has a composition of several gases and is carried out at different pressure levels.
11
The two-cycle C3MR is the dominant liquefaction process for natural gas. It involves a propane cycle
as precooling and then a second cycle of mixed refrigerants. From this principle, Shell has developed
the Shell Dual Mixed Refrigerant (DMR) liquefaction process. This process uses mixed refrigerants in
both cycles and is proposed as a good alternative LNG process on a FPSO [15]. The DMR process has
been selected for the Sakhalin Energy LNG project and is currently under construction. Developments
of the DMR process have further improved process efficiency, from the C3MR process [8]. In addition
and of importance for locations such as the Sakhalin, the DMR process is flexible to various operating
conditions [16]. The site of the Sakhalin plant experience temperatures down to -35 °C in the winter
and 20 °C in the summer [16].
Table 2.2 Efficiency table of expander natural gas liquefiers [4]
The Oman LNG plant is based on propane precooling and mixed refrigerant. It is known as one of the
most efficient liquefiers under operation [4], but the DMR at Sakhalin is expected to have even better
efficiency. Table 2.2 compares the Oman LNG plant efficiency with different expander liquefiers. Due
to its size and complexity it is not preferable for a FPSO but it gives a picture of what is feasible. For
expander processes based on pure refrigerants, the number of cycles has a significant influence on the
efficiency. It has to be noticed that the two dual expander processes have a methane cycle followed by
a nitrogen cycle. Using methane or natural gas cycle instead of nitrogen may give some benefits on
efficiency and will later be explained in Chapter 4.2.
Expander processes have been in focus for floating production with refrigerants in gas phase. They
vary from single to dual cycles involving nitrogen and/or natural gas as refrigerant. Two promising
processes have been developed. Höegh LNG has chosen the NicheLNG process from CB&I and
several companies have proposed a nitrogen dual cycle process.
12
2.1
Niche LNG
The NicheLNG liquefaction process is based on a methane (natural gas) and nitrogen refrigerant cycle,
with one open and one closed cycle respectively. They are independent expansion-compression cycles
but they do overlap each other by heat exchange.
The process has the benefit of operating with high pressures, resulting in smaller pipes and valves than
processes at close to ambient pressure. This gives an advantage related to space constraints on topside
of a FPSO.
Neither of the refrigerants experience a phase change. Both remain in gaseous phase, so there are no
problems with two phase flow distribution. A cycle remaining in gaseous phase reduces the risks of
leakage and eliminates the need for liquid refrigerant storages, drums and separators.
Both advantages of high pressure and no liquid content in refrigerants reduce necessary space and the
equipment count. The process is also more robust with respect to hull movements, due to refrigerants
operating in gas phase. A non-flammable refrigerant, short start-up time, no venting or flaring of
refrigerants after shutdown and smaller footprint increases the benefits of this process. The methane
cycle is a flammable refrigerant and has to be included in the safety evaluation. Nevertheless, the
proposed FPSO-1 already has large amount of liquid hydrocarbons as LPG so this refrigerant cycle
should not have a sufficient involvement in the safety.
Figure 2.1 Flow sheet of the NicheLNG process [5]
Figure 2.1 is a basic principle of the NicheLNG process with two expander-driven compressors, two
compressors and one heat exchanger. It illustrates how the natural gas and nitrogen are utilized as
refrigerants. The open cycle is extracted natural gas to be expanded and then re-enters the heat
exchanger. This cycle is known as a Claude cycle. In addition, a closed nitrogen cycle cools in the
same temperature range but takes first care of the subcooling. The flash gas from boil-off is not shown
in Figure 2.1.
13
The specific energy consumption is estimated to 16.5 kW/tonLNG/day (0,396 kWh/kgLNG) [4]. A
relatively low energy consumption when compared with other processes based on pure refrigerants, as
expressed in Table 2.2. The Dual Expander C1/N2 corresponds to the NicheLNG process.
2.2 PRICO - Single Mixed Refrigerant
This is one of the most basic processes based on mixed refrigerant and has a low equipment count.
The setup is one heat exchanger network with a mixed refrigerant consisting of methane, ethane,
propane, pentane and nitrogen. The composition is chosen based on the respective boiling points of the
components to match the mixed refrigerant curve with the cooling curve of natural gas. Closer curves
will increase the efficiency.
Figure 2.2 Basic principle of a single mixed refrigerant process [7]
2.2.1 Principle
A mixed refrigerant containing different gases is pressurized through a compressor. The discharge
pressure has to be sufficiently high to give enough cooling duty after a later expansion. Mixed
refrigerant flows through the heat exchanger as a hot stream. After leaving the heat exchanger in the
cold end, the refrigerant undergoes a pressure decrease through an isenthalpic expansion valve. The
reduction in pressure and temperature, by heat exchange and expansion, is necessary to achieve
enough cooling duty. The stream contains now liquids. Then it reenters the heat exchanger as a cold
stream. Heat transfer from the two hot streams, natural gas and mixed refrigerant, evaporates the liquid
over a wide temperature range. The pressure is then recovered by compression.
Natural gas to be liquefied has initially a higher pressure level than the ambient condition. Not shown
in figure 2.2 is a valve in the cold end. After heat is released in the heat exchanger an expansion brings
the natural gas to the specifications required of LNG.
2.2.2 Extensions of PRICO
This process has a considerably large flow rate of refrigerant which leads to high compression work.
On the other hand, the necessary pressure ratio is lower than for an expander liquefaction process as
the NicheLNG. An improvement is compressing over two stages with inter-cooling reduces the
consumption of work.
Also, the pressurized natural gas to be liquefied has a potential of work recovery. By utilizing the
pressure to lower the temperature, through a turbine, work and cooling duty will be produced. This
concept is the same as the open methane cycle in the NicheLNG process. The work recovered from the
turbine can be utilized with a generator or a directly driven compressor.
14
2.3
Dual nitrogen refrigerant
Some proposed solutions for offshore LNG production have been with nitrogen as refrigerant. A
nitrogen based liquefaction application has some advantages over the other compact LNG processes
intended for offshore environment. It is easier to model, and the equipment is easier to operate,
because the nitrogen refrigerant is always in gas phase. Due to the gas phase and the fact that nitrogen
is an inert gas, the process is safer because of reduced hydrocarbon inventory compared to other
processes.
The setup of the different nitrogen liquefaction processes share specifications as operating at high
pressure levels and normally two refrigerant cycles. They do have some differences as the outlet
pressure of the expanders. This will affect the size due to suction volume and complexity of the
process. Illustrations and discussion of the possibilities will be covered later in this thesis.
Figure 2.3a Statoil proposed solution
Figure 2.3b BHP Billiton proposed solution
Figure 2.3a illustrates the Statoil solution with pressure reduction to two levels. A large amount of the
refrigerant stream flows through the middle pressure expander. This stream is supposed to cover the
refrigeration of precooling and liquefaction. The low pressure stream covers the subcooling and the
rest of the liquefaction and precooling. The flexibility of the process is rather small and is limited to
the temperature splits between the cooling stages and the compressor pressure level.
In figure 2.3b the expanders have the same pressure reduction, and the two cold streams meets and
flow as one cold stream through the heat exchanger from the subcooling liquefaction split. The most
important factors from the two examples are the mcp variations through the heat exchangers and the
suction volumes of the compressors.
15
2.4
Comparing conventional with expander liquefaction processes
Focusing on efficiency there is no doubt that mixed refrigerant processes have a better efficiency than
pure refrigerant expander processes [4]. Expander processes have lower efficiency but many benefits
when production is in an offshore environment. The following sections compares selection criteria of
different processes.
Compact:
Mixed refrigerant cycles require large storage capacity [4]. The large flow rate takes up area and
increases the weight. Heat exchangers and equipments have to be able to operate with two phase flow.
Gaseous refrigerants, such as nitrogen, have the potential of being compact because there is no
refrigerant storage and the refrigerants are operating with high pressures. Although the refrigerant flow
rates are decreased, the required heat transfer area may not decrease because of the refrigerant heat
transfer coefficient is also much lower. Non-flammable refrigerant will also reduce necessary area for
safety.
Safety:
Operating with flammable refrigerants is well known from earlier LNG plants. Even though these
plants have good safety records, operating on a ship with restricted area gives stricter safety concerns.
Mixed refrigerant and cascade processes have large flammable refrigerant inventories, high circulation
rate and flare requirements.
An expander process with nitrogen as refrigerant has higher inherent safety because nitrogen is an
inert gas. As for the NicheLNG operating with natural gas as refrigerant, some stricter safety issues are
introduced.
Operation:
Mixed refrigerant processes have a more complex operation due to refrigerant composition and high
equipment count. It has also a longer start-up time and flare requirements.
Expander processes have an advantage in all three of the following process selection criteria: Ease of
operation, quick start-up time and low equipment count.
Efficiency:
Fluids going through vaporization have to attract heat under almost constant pressure. This change of
state characterizes a typical single composite refrigerant process. A mix of fluids with different boiling
points flowing through a heat exchanger results in an evaporation of the cold stream. As Figure 1.8
shows, the natural gas curve does not have a linear profile. With the right composition of gases in the
refrigerant, a gliding temperature profile is possible.
Expander processes operate in gas phase. The heat transfer cannot benefit from evaporation. For pure
gases the specific heat is almost constant so a variation in the refrigeration flow rate is necessary to
cover the non-linear temperature profile of natural gas.
16
3
Promising liquefaction processes for FPSO applications
In the proposed FPSO-1 after the pretreatment of the natural gas and after the LPG separation lean gas
enters the liquefaction section. This section consists of two identical trains which have a total
production rate of 4670 ton/day of LNG. The feed stream is divided into two equal mass flows
entering the two trains.
3.1
The NicheLNG process
Figure 3.1 Pressure – temperature diagram illustrating the natural gas path [5]
As shown in the pressure-temperature diagram in Figure 3.1, recycled natural gas mixed with the feed
gas is compressed to a pressure above cricondenbar (1-2). An aftercooler, utilizing water as
refrigerant, will then lower the temperature (2-3). Then natural gas is entering the main LNG heat
exchanger as a hot stream, and is cooled against cold low pressure natural gas and a nitrogen stream
(3-4). Before further cooling, an amount of the hot natural gas stream is extracted from the heat
exchanger and sent to an expander. The pressure of the extracted natural gas is reduced by a turbine
(4-5). This extracted natural gas stream now acts as a cold stream. Together with the nitrogen stream,
heat is now removed from the pressurized natural gas stream (5-6). Exiting the heat exchanger natural
gas flows to a compressor which is mounted on the same shaft as the expander. Energy generated from
the expander is utilized by compressing the natural gas. The discharge from the compressor is cooled
by an after cooler (6-1).
The remaining natural gas to be liquefied, which has the same flow rate as the feed, is further cooled
(4-8). Pressure is reduced across a valve which results in entering the two phase region and produces
some flash gas and LNG (8-9).
17
Figure 3.2 Pressure – temperature diagram illustrating the nitrogen path [5]
To bring the natural gas to the required temperature (state 8), nitrogen is used as a second refrigerant
in a closed loop. The warm nitrogen stream leaving the heat exchanger is compressed followed by an
aftercooler and then further compressed by an interstage compressor with an after cooler (1-2). It then
flows as a warm stream through the heat exchanger and is cooled by the cold streams of natural gas
and nitrogen (2-3). Then the high pressured nitrogen stream is expanded through a turbine and
produces cooling and work (3-4). The expanded nitrogen provides a cooling potential at low
temperature that is utilized in the heat exchanger (nitrogen is heated from state 4-1).
Figure 3.3 Temperature profile in the heat exchanger
18
The temperature-enthalpy diagram in Figure 3.3 shows the temperature difference in the heat
exchanger. As illustrated in Figure 3.3, the pinch point is in the warm end. The cold composite curve,
with temperature split at -85°C, has an almost linear profile in two intervals.
The natural gas composition in the 2DLE case is chosen after the pretreatment and the LPG separation
to meet the LNG specifications. Some processes have an integrated LPG extraction. A partial or full
integration depends on feed gas and onshore/offshore production. Since the focus is on the NicheLNG
process the natural gas composition from the 2DLE case applies for all simulated cases in this thesis.
3.1.1 Exergy analysis of the NicheLNG process
One way of analyze the quality of the NicheLNG process is by determining the exergy efficiency.
From the ratio of minimum work of liquefaction and the actual compressor work, the exergy
efficiency can be expressed. The exergy calculations are done with enthalpy and entropy values
obtained from simulations in HYSYS.
Since exergy is dependent on pressure, temperature and ambient conditions, the initial state will affect
the efficiency. The feed gas alone has a relatively high exergy value. This affects the exergy efficiency
and the specific power consumption. Hence, comparison of processes based on overall exergy
efficiency and specific work will favor the one with highest inlet feed pressure.
Figure 3.4 Specific work as function of overall exergy efficiency for a LNG process from gaseous feed at different
pressures to saturated liquid at 1 bar
As Figure 3.4 illustrates, the minimum work (exergy efficiency of 100%) for liquefaction of natural
gas at atmospheric pressure (1 bar) is 0,278 kWh/kgLNG. The calculations are done with enthalpy and
entropy values from a feed gas simulated in HYSYS, and are attached in Appendix A. Included in
19
Figure 3.4 is feed gas at higher pressures. The exergy content increases with higher pressure and
reduces necessary specific work at a given overall exergy efficiency. So when efficiency of the
NicheLNG process is to be compared with other liquefaction processes, the operating conditions have
to be uniform. The efficiencies given in Table 2.2 are not specified with operating conditions so they
may not be comparable. The efficiency of Dual Expander C1/N2 (NicheLNG) in Table 2.2, do not
match with the efficiency 0,503 kWh/kgLNG, simulated in HYSYS by the 2DLE file. An explanation of
the variation can be the composition of feed gas, cooling water temperature, chosen equipment and the
condition of feed gas after pretreatment. To be able to compare the NicheLNG process with another
liquefaction process, a simulation of a dual nitrogen expander process has been done at equal
conditions. The dual nitrogen expander process will be investigated later.
In the matter of comparing the NicheLNG process with another process, some design specifications
have to be set in the evaluation. The main compressor use vendor curves in the HYSYS 2DLE file. So
to achieve equal conditions for the two concepts to be investigated, all compressors were defined with
polytropic efficiency of 82% and expanders with adiabatic efficiency of 87%. Every compressor has
an aftercooler which lowers the temperature of the compressed gas to 38°C. The minimum
temperature approach in the heat exchanger is specified to 3°C. Since phase change occurs in the heat
exchanger, the ‘Weighted model’ was chosen as heat exchanger parameter for UA-value calculation.
The background for the chosen model is attached in Appendix B. Exergy calculations are done with
ambient conditions at 1 bar
and 25°C.
From the specifications above
the NicheLNG process is
simulated in HYSYS. The
transition from compressor
vendor curves resulted in a
lower outlet pressure of the
expander driven compressor.
To obtain feed gas pressure
of the open methane cycle
was an additional compressor
(NG Comp.) installed to
increase the pressure to the
same pressure as the feed gas.
Table 3.1 Efficiency and work consumption for the NicheLNG process
The minimum power for
liquefaction was calculated with values from feed gas conditions to after
the heat exchanger, as discussed in Section 1.3. Feed gas composition
from Table 1.2 has been used for all cases.
With feed pressure at 30 bar, the exergy efficiency was calculated from
Eq. 1.11 to be 31,2%. As can be seen from Figure 3.4 the calculated
exergy efficiency matches with the calculated specific work in Table 3.1.
A way of characterizing the quality of a liquefaction process is by
comparing it with the theoretical minimum liquefaction work. Table 3.2
expresses necessary minimum work for given pressures. The calculations Table 3.2 Minimum
are done with data from HYSYS. The calculations resulting in Table 3.2 liquefaction work to feed gas
pressure
20
are found in Appendix A and the composition of the natural gas is found in Table 1.2. Table 3.2
illustrates how feed gas pressure affects the efficiency and will be further investigated in Chapter 4.
Figure 3.5 The composite curves for the NicheLNG process
Figure 3.6 Flow sheet of the simulated NicheLNG process
The N2 Comp. in Figure 3.6 represents two compressors with interstage cooling.
21
3.2
Dual nitrogen process
The most frequently proposed process for offshore liquefaction of natural gas is with two nitrogen
refrigerant cycles. One cycle covers the precooling and the second is used for liquefaction and
subcooling. Nitrogen as refrigerant medium is very flexible. The relative low dew point gives the
opportunity of the refrigerant cycles to operate at a wide temperature range. Chapter 2.3 describes
different concepts. The choice of concept in this work was with emphasis on equipment count that
matches the NicheLNG process.
3.2.1 Exergy analysis of a dual nitrogen process
To compare a dual nitrogen process with the NicheLNG process equal conditions are necessary. The
same efficiencies for compressors and expanders are being used. Feed gas and the produced LNG have
the compositions from Table 1.2 and the same conditions as in the exergy analysis for the NicheLNG
process. The natural gas in the open methane cycle has been replaced with 98 mole% N2 and 2 mole%
O2 and configured to a closed cycle. With these specifications, flows and pressures in the two nitrogen
cycles have been optimized to minimize work consumption. Figure 3.8 illustrates the design of the
dual nitrogen process.
Table 3.3 expresses the
quality of the simulated dual
nitrogen process. When the
natural gas leaves the cold
box, a pressure reduction
through a valve will result in
flash
gas
formed
by
evaporation. The evaporation
is necessary in order to
reduce the nitrogen content
to 1 mole%. In both
processes, the natural gas
leaving the cold box has the
same temperature at -157,2
°C. With temperatures at the
same level the flash gas Table 3.3 Efficiency and work consumption for the dual nitrogen process
production will be at the same
rate and thereby a similar LNG composition. Both processes produce LNG with a higher heating value
of 10,95 kWh/m3 (1058 BTU/scf).
It was decided to operate the refrigerant with only one high pressure and one low pressure level,
respectively at 90 bar and at 17 bar. Then only adjustment of the flow rate was necessary to provide
the required cooling.
22
Figure 3.7 The composite curves for the dual nitrogen process
Figure 3.8 Flow sheet of the simulated dual nitrogen process
The N2 Comp-1 in figure 3.8 represents two compressors with interstage cooling.
23
3.3
Discussion of NicheLNG versus dual N2 process
Both processes are almost similar in design but are distinguished by the different refrigerant. They
have the same amount of compressors and expanders. Equal equipment count is one criterion to
achieve a fair comparison. In addition to the two expander driven compressors had both processes four
compressors.
The simulated NicheLNG process in Section 3.1.1 distinguished from the original 2DLE by adding
one compressor in the open methane cycle. This had to be done after efficiency adjustment of
compressors and heat exchanger. It looks like that 2DLE low pressure of the open methane cycle was
defined by the feed gas pressure or vice versa. A design with an open methane cycle is influenced by
the feed gas pressure. If the feed gas pressure changes it will affect the cooling duty of the open
methane cycle, since feed gas and refrigeration gas are pressurized by the same compressor. By
closing the loop the NicheLNG process will be more available to different feed pressures. It may also
be an advantage to have a possibility of higher pressures in the methane cycle.
The NicheLNG process has about 10 % lower work consumption. This can be explained by the chosen
refrigerant and the heat distribution in the heat exchanger. Natural gas as a refrigerant has higher cp
than nitrogen, resulting in a significantly smaller refrigerant mass flow rate. Higher mass flow rate
causes more irreversibilities in the compressors and expanders, so the lower mass flow rate explains
why NicheLNG is the most energy efficient process.
Table 3.4 Results of both processes
Figure 3.5 and 3.7 illustrates that the processes do not differ much in temperature difference and they
have almost the same UA value. A larger temperature difference produces more entropy. With closer
composite curves (smaller ∆T), the process efficiency will be improved, however, at the expense of a
larger heat exchanger.
The NicheLNG uses a flammable refrigerant, so it has a disadvantage from a safety point of view. An
evaluation of this issue must be considered with respect to the safety requirements.
24
4
Adjustments and analysis of NicheLNG
4.1
Precooling
To keep refrigerant cycles at relatively small sizes, even at high LNG production rate, the number of
refrigerant cycles can be extended. A precooling helps the more energy demanding liquefaction and
subcooling by reduced flow rate of refrigerants. In addition, the specific heat of natural gas varies with
temperature so refrigerants with different mixtures can benefit from closer temperature difference in
the heat exchangers.
The most common precooling refrigerants are propane or ethane/propane mixtures. These are
flammable hydrocarbons and have to be included in the evaluation of the process safety. From Table
2.2 a precooling integration can provide an 18% higher efficiency to the NicheLNG process. Another
possibility is CO2 as precooling refrigerant. CO2 is not flammable and may be preferable in offshore
liquefaction processes. A disadvantage is the risk of forming solids.
Figure 4.1 The NicheLNG process with a precooler in front [17]
Figure 4.1 is an illustration of a precooler installed for a single train. The FPSO-1 is planned to have
two liquefaction trains, so a larger precooler to cover both will keep equipment count down.
25
4.2
Refrigerant medium
All of the expander processes described in Chapter 3 operates with two refrigerant cycles based on
nitrogen or natural gas and nitrogen. They operate in gaseous phase through the cycles and have
therefore some constraints. An expander process has to take into the account the boiling point of the
gas, due to problems with liquids in a turbine. Hence, a natural gas refrigerant cycle has a stricter
constraint on operability than nitrogen. Since the boiling point of nitrogen is -196 °C at atmospheric
pressure, entering the two phase region is at a lower risk than operating with natural gas. The relative
low boiling point results in a higher degree of freedom than for a refrigerant of natural gas.
Figure 4.2 Simple illustration of expander precooling
Figure 4.2 is an example of a closed precooling cycle. It contains compressor, expander, cooler and
heat exchanger. Simulation is done with an adiabatic efficiency of 80% in the compressor and
expander and no pressure loss the through aftercooler and heat exchanger. The heat exchanger has a
minimum temperature approach of 3°C.
The calculations shown in Table 4.1 are
done with a constant pressure ratio of
refrigerant and with a desired outlet
temperature of the natural gas. Mass flow
is adjusted to cool down the natural gas to
-65°C. The chosen refrigerants are pure
methane and nitrogen.
This example shows the importance of the
refrigeration gas ability to extract heat.
Methane as a refrigerant has a higher
specific heat capacity, hence lower
necessary mass flow resulting in higher
efficiency. A nitrogen refrigeration cycle
has, with the specifications from Table 4.1, Table 4.1 Specifications and results of a precooling example
over twice the energy consumption as a
methane refrigeration cycle.
Ethane has an even higher specific heat capacity. So for the NicheLNG process, operating with an
open refrigeration cycle of natural gas, small amounts of heavier hydrocarbons such as ethane will
give a small increase in efficiency.
26
4.2.1 Change in gas characteristics for pressure variations
The composition of a stream flowing through a heat exchanger affects the size. The amount of heat a
fluid is able to hold depends on its specific heat capacity and flow rate. If a fluid is supposed to attract
a given heat duty at a given temperature range the flow rate will be dependent on the specific heat
capacity. From this a higher specific heat capacity will result in a lower necessary flow rate, and
thereby influence the heat exchanger size and the compressor work.
Figure 4.3 Specific heat capacities of N2 and CH4 at pressure levels from the NicheLNG process
Figure 4.3 illustrates the variation of specific heat capacity at constant pressure levels between two
gases in a certain temperature range. The most important information from Figure 4.3 is the large cp of
methane. Even though methane is able to attract more heat than nitrogen is its range of operation in
gaseous phase restricted to a smaller temperature region. The peaks for each curve in figure 4.2 are the
critical points of the respective gases.
It is easy to conclude that CH4 is able to hold more heat than N2 at the same flow rate. Hence, it is
believed that the choice of refrigerant will affect the heat exchanger size. It has to be noticed that this
is not fully true since the size of heat exchangers are also dependent on the heat transfer coefficient. At
a fixed conduction through the wall of a heat exchanger the heat transfer coefficient is relative to
convection. Speed will vary with the chosen gas to satisfy necessary heat transfer, and thereby
influence the turbulence with change in speed and viscosity.
27
The overall specific efficiency of a liquefaction process is the ratio between the work consumed and
the LNG produced. To get necessary cooling duty, a gas has to be expanded and later re-compressed.
The compression work of a single compressor is depending on inlet temperature, pressure ratio and
specific volume of the gas. Compression in liquefaction processes is done at close to ambient
temperature. The pressure ratio depends on desired cooling duty which is also influenced by the
chosen refrigerant gas.
Figure 4.4 Specific volume variations with pressure at a temperature of 30 °C
Figure 4.4 illustrates how the specific volume varies with pressure for different gases. The sudden
decrease of the propane (C3H8) above 1000 kPa is due to condensation.
The molecular weight of gases influences its specific volume. For natural gas containing mostly
methane a small amount of heavier hydrocarbons will give a reduction in specific volume when
comparing with pure methane. Nevertheless, it will not decrease below the specific volume of
nitrogen.
Maybe the most important observation from Figure 4.4 is the specific volume value at higher
pressures. The specific volume difference between certain gases reduces with higher pressure. So
operating at high pressure levels will not give significant difference in the specific volume. The
specific heat capacity will be more effective. A high specific heat capacity demands lower flow rate
for a given heat exchange duty.
28
4.3 Placement of expansion
As explained earlier an expansion instead of a throttling is favorable. In sub-ambient processes can a
pressurized stream provide cooling and produce work. A typical expander liquefaction process is
normally based on two refrigeration cycles. Each cycle is supposed to cover a given temperature
region.
Figure 4.5 The NicheLNG open methane refrigeration cycle with refrigeration regions indicated
Figure 4.5 illustrates the open methane cycle in the NicheLNG liquefaction process. It has the same
principle as a typical Brayton cycle. The placement of the expander is at -1,5°C after an internal heat
exchange. After the expansion to 19,5 bar the temperature is reduced to -76,8°C which results in a
refrigeration load of 2568 kJ/kgmole and a necessary compressor work of 2845 kJ/kgmole (17084
kW). The calculations are done with the natural gas refrigerant flow rate from the 2DLE case at
21618,85 kgmole/h.
4.3.1 Expander placement for the open methane cycle
A way of expressing the efficiency of the open methane cycle is by the ratio of compression work and
refrigeration load. By comparing the original design with other hypothetical inlet temperatures the
quality can be found. It has to be noticed that the results from Figure 4.6 is restricted to the open
natural gas cycle in the NicheLNG process. How the overall efficiency varies with different inlet
temperature of the open cycle expander is not included.
29
Figure 4.6 Efficiency of the open natural gas cycle with different inlet temperatures
Figure 4.6 illustrates how the efficiency of the open
methane cycle varies with inlet temperature from the
original indicated with a dotted line. The calculations
are attached in Appendix C. All calculations are done
with an ideal heat exchanger, turbine with 87%
adiabatic efficiency and compressors with 82%
polytropic efficiency. The ∆T in the warm end is set to
2°C.
From Figure 4.6 and Table 4.2 less work per
refrigeration load will be the result with cooling down
to -10°C before an expansion. As earlier mentioned,
these calculations are restricted to the open natural gas
cycle. More internal heat exchange, with inlet
expansion temperature at -10°C, will give less
refrigeration duty to the overall process.
Table 4.2 Results of inlet temperature as
original at -1,5°C and at -10°C
4.3.2 Consequences of expander
placement
Inlet expansion of the open methane cycle at -10°C results in almost the same cooling duty (internal
heat exchange and refrigeration duty) as with an inlet expansion temperature at -1,5°C. Hence, less
work is needed for the same cooling duty. Even though the total cooling produced does not change,
the natural gas has been further cooled by internal heat exchange and this will affect the refrigeration
load. The reduced refrigeration load is 473,6 kJ/kgmole or in heat flow 10,239*106 kJ/h. This reduced
refrigeration load has to be covered by the closed nitrogen cycle or by higher compressor work in the
open methane cycle.
30
4.4
Liquefaction pressure
Natural gas to be liquefied is always under high pressure. The advantage of liquefying pressurized
natural gas can easily be seen from a pressure - enthalpy diagram. Natural gas at constant temperature
has lower enthalpy with increasing pressure. Hence, the amount of heat to be removed is reduced with
increased pressure at a constant temperature. After subcooling, the natural gas is still pressurized. An
expansion, in addition to recovering work and contribute to the cooling, will bring the natural gas to
the given specifications.
Figure 4.7 Natural gas path through liquefaction for a typically onshore facility [11]
Figure 4.7 illustrates the natural gas path through liquefaction for a typical onshore based LNG
production. The composition is C1 89,7%, C2 5,5%, C3 1,8% and N2 2,8% [11]. The natural gas
enters with a pressure of 60 bar. As figure 4.7 illustrates is the cooling done below the critical point.
By entering the two phase region some liquid can be extracted with a phase separator before further
cooling. The NicheLNG process avoids the utilization of a phase separator by higher liquefaction
pressure. How the liquefaction pressure affects the liquefaction will be further investigated in the
following sections.
4.4.1 Relationship between feed gas and liquefaction pressure
When production of LNG is done offshore some constraints will affect the process. By excluding a
phase separator, the pressure can be increased above the critical point. Without a phase separator, the
natural gas entering has the same composition as the one leaving the cold box. A lean composition is
therefore required to meet the LNG specifications. This will affect the comparison of efficiency
between different LNG processes. An integration of a phase separator in the liquefaction section will
affect the power consumption due to a different liquefaction pressure. For the NicheLNG process, all
LPG fractionation is done in front of the liquefaction. For the processes in Table 2.2 a detailed
description on the utilization of phase separator in the liquefaction part could not be found. So a direct
comparison of a traditional onshore process as Oman LNG cannot be made without more data.
31
As cooling water is relatively cheap there are several advantages by using this as a cooling medium.
When it comes to saving work, two areas in particular are outstanding. Multiple compressions with
interstage cooling in order to achieve close to isothermal compression will save some energy. Second
the cooling water temperature
is important for the amount of
necessary heat to be removed.
Figure 4.8 illustrates an ideal
liquefaction
process
for
natural gas. The ideal work is
shown in the W area with heat
rejection
at
constant
temperature. The T-s diagram
in figure 4.8 shows how the
influence of the heat rejection
temperature has to work
consumption.
Heat is removed as the gas is
cooled at gliding temperature
and constant pressure. The
isobar lines show two
important
factors
with
increased
pressure.
As
pressure increases the gliding
temperature gets more linear.
A more linear cooling of the
Figure 4.8 Ideal liquefaction process of natural gas [11]
natural gas is an advantage
when the number of refrigeration cycles is restricted to one or two. The more interesting factor is the
changing properties of the feed gas with increased pressure. Figure 4.7 illustrates the enthalpy
reduction with increased pressure at a constant heat rejection temperature. A lower enthalpy of the
natural gas entering the cold box results in less heat to be removed, and thus lower necessary work
consumption for the refrigeration cycles. Even though compression of the feed gas consumes more
work with higher pressure, it may be favorable due to the low refrigeration efficiency for liquefaction
processes.
Figure 4.9 Simple flow sheet of the liquefaction path from feed gas to LNG
Figure 4.9 illustrates the liquefaction path from feed gas (a) to LNG (c). A compression of the feed gas
(a-b) will change its properties which will influence the liquefaction. In order to illustrate how the feed
gas pressure affects the liquefaction some adjustments to the 2DLE process have been made. The
32
original process is the 2DLE with one open methane refrigeration cycle. Closing the loop and using a
feed gas compressor will simplify the analysis by focusing only on the natural gas to be liquefied. This
adjustment will also give the possibility of limiting the boundaries to the liquefaction part only (b-c).
The adjusted process will be referred to as 2DLE-2. Values given in Table 4.3 are from the simulated
2DLE-2 and are based on the principle from Figure 4.9. Stream b is after a compression and
aftercooler respectively at 75 bar and 38°C. The feed gas goes through liquefaction and ends up as
stream c after the cold box at 74 bar and -157,3°C. The pressure drop of 1 bar is losses through the
heat exchanger.
Table 4.3 Exergy calculations of the liquefaction part (without feed gas compressor and for a single train)
Liquefaction efficiency is calculated after the feed gas compressor. In Table 3.1 the feed gas
compressor is included when efficiency is calculated. A reduction in efficiency will therefore occur for
the liquefaction part, b-c in Figure 4.9, due to the relative high efficiency of compressors.
The high efficiency of compression and the low efficiency of liquefaction make the change in
properties interesting when pressure is increased. Figure 4.7 illustrates that when pressure is increased
at constant temperature the enthalpy reduces. A lower enthalpy results in lower refrigeration work.
Figure 4.10 Specific work of compression from 1 bar to a certain liquefaction pressure and specific liquefaction work
33
To benefit from high liquefaction pressure, a compressor is necessary to reach the desired pressure.
Figure 4.10 illustrates work consumption for compression alone and compression included
liquefaction work. Natural gas is compressed from 1 bar with composition from Table 1.2 and then
refrigerated from a certain pressure at 38°C to -157,3°C. The feed gas compressor has a polytropic
efficiency of 82% and the liquefaction part is illustrated with different efficiencies. The data are given
in Appendix D. Figure 4.10 illustrates that only liquefiers with low efficiencies will benefit of higher
liquefaction pressures when the natural gas enters with a pressure of 1 bar. If natural gas, after a
pretreatment of LPG, has a pressure higher than 1 bar a liquefier with higher efficiency will benefit
from higher liquefaction pressures. Hence, the influence of the feed gas compressor work will be
smaller on the whole process.
Figure 4.11 Work consumption for compression from a feed gas pressure to a liquefaction pressure and a 26,6%
efficient liquefier
In Figure 4.11 feed gas at different pressures is illustrated. As illustrated a 26,6% efficient liquefier
will benefit from higher liquefaction pressure for feed gas pressures at a range from 10 to 60 bar. The
NicheLNG, expressed in blue, has a feed gas pressure after pretreatment of 30 bar. The total work
consumption in NicheLNG is reduced with higher liquefaction pressure. Figure 4.11 is based on
calculations attached in Appendix E-2.
If some adjustment to the liquefier is done to improve its efficiency new calculations are necessary. A
higher efficiency for the liquefier will influence the benefits of higher liquefaction pressure. The
compressor work will dominate more of the total consumption with higher liquefaction pressure.
Appendix E-2 shows how an increase in efficiency of the liquefier will give a negative influence to the
overall process, when liquefaction pressure is increased.
34
4.5
Discussion on the analysis
A decision of operating with high pressures in the process keeps the specific volume down. Low
specific volume is beneficial to equipment sizes. Smaller equipments can be chosen and therefore
lower necessary space for the plant. When operating at high pressures, the specific volume is less
important when selecting the refrigerant gas. The difference in specific volume between nitrogen and
hydrocarbon gases is very small with higher pressures. More important is their respective heat
capacity. It is shown that methane has a relatively high specific heat capacity. In addition, the specific
heat capacity of methane is more influenced by temperature variation. At high pressures, the specific
heat capacity of methane increases with a reduction in temperature. The benefits of high heat capacity
relates to lower necessary mass flow and thereby lower compressor work. So when expanders are used
in a refrigeration process, methane rather than nitrogen, is more effective. It has to be noticed that
methane has some restrictions on operability due to higher dew point.
The open methane cycle in the NicheLNG process enter the cold box at 38°C and cools down to 1,5°C before an expansion. By cooling it further down to -10°C before expanding, less work is
necessary for the same cooling duty. How it will affect the whole process is not investigated, so this is
not something one can conclude.
An increase in pressure of natural gas results in higher exergy. Cooling water is a relatively cheap
refrigeration resource and is therefore seen as free. By cooling a pressurized natural gas to be liquefied
with cooling water an increase in pressure will result in lower enthalpy. The stream leaving the cold
box has relative high exergy and an increase in exergy of the natural gas to be liquefied results in a
smaller necessary exergy change. Hence, less work is required to remove heat.
When evaluating LNG processes with emphasis on how the liquefaction pressure influence the process
the control boundary is of importance. The state of the feed gas affects the benefits of higher
liquefaction pressure. A low feed gas pressure with a high efficiency liquefier (40 – 50% or higher)
may have a negative effect on the total work consumption if the liquefaction pressure is increased. For
the NicheLNG process and other offshore based liquefiers, the efficiencies are low. As shown in
Chapter 4 and Appendix D-2 liquefiers with low efficiency will take advantage of higher liquefaction
pressure. The NicheLNG has a liquefaction pressure of 75 bar and a liquefier efficiency of 26,6%.
With these characteristics an increase in liquefaction pressure has positive effect for any entering feed
gas pressure.
35
36
5
Increased capacity of the NicheLNG process
Destination and customer for the FPSO is not yet decided. In order to have a more flexible design to
meet future demands on production rate, Höegh LNG wants to look at the possibilities to expand LNG
production rate of the original design. An increased capacity with implementation of additional units
will be covered in this chapter. Focusing on the efficiency for each improvement, an indication of
increased capacity can be shown.
The selected improvements to be investigated:
•
•
•
•
Utilization of end flash gas (EFG)
Liquid expander
Two stage compression
Higher UA value
The inert gas nitrogen affects the higher heating value of the LNG. Higher nitrogen content in LNG
results in reduced heating value. Nitrogen is a more volatile gas than hydrocarbons so the EFG leaving
the LNG receiver contains more nitrogen than the LNG. The EFG production is therefore necessary to
obtain desired LNG specifications. All simulations in Chapter 5 are done with same efficiencies and
UA-values as in the 2DLE case.
5.1
Utilization of end flash gas (EFG)
The cold duty from EFG leaving the LNG receiver is in the original design not utilized. It constitutes
about 7% of the feed mass flow rate to be liquefied but has only available 3% cold duty of the
necessary cooling of the feed gas. A certain amount of EFG has to be produced in order to meet the
specification of LNG. Work consumed increases with the rate of EFG production. Due to the low
temperature of EFG (-162°C) the work to produce this cold duty is relative high, referred to figure 1.3.
Hence, the EFG has a relative high quality. From Eq. 1.6 the COPth,max is calculated to 0,555. A COP
at 0,555 and EFG cold duty at 739 kW results in a work input of 1331 kW. The calculated COP in this
case is when operating reversibly and adiabatically. To achieve the same cold duty in a real process
higher work consumption is
necessary.
With the real COP for the 2DLE
case a calculated efficiency increase
of including EFG could be
controlled. When including EFG as
a cold stream in the cold box, the
overall work consumed is simulated
in HYSYS to be 47785 kW with the
same production rate of LNG. From
Table 5.1 the required real work to
obtain the cold duty of EFG is 1677
kW. Adding the calculated work of
EFG cold duty to the simulated
work of the 2DLE with EFG, results
in a power consumption of 49462 Table 5.1 COP of 2DLE and the EFG
kW which is close to the power
37
consumption in the 2DLE case. Some losses in the cold box may explain the small difference between
the calculated and the simulated power saved with integration of EFG.
5.2
Liquid expander
As the natural gas to be liquefied is under a relatively high pressure and in liquid state when leaving
the cold box, a liquid expander may give some improvements to the process. In the 2DLE case,
pressure reduction is done through an isenthalpic valve. An isenthalpic expansion results in no work
recovery and has a smaller temperature drop than for an isentropic expansion. By installing a liquid
expander at the cold side of the cold box, work can be extracted and required cooling duty of the heat
exchanger may be reduced.
In order to avoid destruction of the pressure exergy a liquid expander was introduced. This expander is
essentially a pump run backwards that allow a subcooled liquid to be isentropically expanded almost
to its bubble point. The most important benefit is the temperature reduction at very low temperature as
power recovery is small at very low temperatures. Hence, isentropic expansion is an efficient way of
rejecting heat, and not necessary as work recovery.
Two-phase expanders are now available and will contribute to fulfill an isentropic expansion into the
two-phase region [10]. These expanders are not proven at large scale and will therefore not be covered
in this thesis.
Figure 5.1 Pressure-temperature diagram for pressure reduction with valve and liquid expander followed by a valve
Figure 5.1 illustrates the cooling path of natural gas to be liquefied. The green line indicates the 2DLE
case and the red is after installation of a liquid expander. Simulated in HYSYS both processes produce
the same amount of LNG, 2357 ton/day for one train. As the temperature out of the heat exchanger is
higher with a liquid expander, less refrigerant duty is required. Compressor power can then be reduced
38
either in the methane or nitrogen cycle. To express the gained efficiency by introducing a liquid
expander, the nitrogen refrigeration flow rate was reduced. This resulted in 2,155 MW reduced total
power consumption and represents 4,3% of the 2DLE case.
Another graphical illustration of how a liquid expander contributes to the efficiency is by a
temperature-entropy diagram. In Figure 5.2, a natural gas at 75 bar and 38°C is cooled down to 157,2°C.
Figure 5.2 A graphical overview work saved by integration of a liquid expander
In Figure 5.2 the heat Q is extracted from the natural gas as a heat source and rejected to the
surroundings as a heat sink. The necessary work W and Wsaved are represented respectively in the blue
and green area above Q. With a liquid expander the amount of work is reduced to involve only the W
area.
Figure 5.2 illustrates that at low temperatures, higher outlet temperature of the heat exchanger result in
a considerable work reduction. By introducing a liquid expander the COP of the reduced duty can be
calculated and be an indication of saved work with increased LNG production.
39
Table 5.2 Results from the integration of a liquid expander
In Table 5.2 the values for both 2DLE cases are simulated in HYSYS. Based on the two cases the
COP is calculated and can then be used as an indication of reduced work with liquid expander at
higher LNG production. Even though reduced work increases with LNG rate, reduced work in
percentage will stay constant.
5.3
Two stage compression
The compression of feed gas and methane in the refrigeration cycle is done with one compressor. The
flow rates of these streams are 63% of the total flow rate (natural gas and nitrogen) through the cold
box and require a 22,3 MW compressor. By including a second compressor in series with an
intercooler, power consumption will be reduced.
The new compressor is assumed to have the same efficiency as the one already in place. To find the
optimum pressure increase for the first compressor, a case was simulated in HYSYS. With a pressure
variation from inlet pressure to liquefaction pressure the optimum middle pressure where found to be
48 bar. This resulted in a power reduction of 1,56 MW and represent a reduction of 6,8%. For the
2DLE process, a two stage compression reduced the overall power consumption by 3,2% at the same
LNG production.
40
5.4
The improvements influence by higher LNG production
Höegh LNG wants to look at how an increase in production affects the process and how new
improvements will contribute to keep the power consumption down. The three proposed solutions to
increase the efficiency of the NicheLNG are all promising without huge changes to its original design.
Additional units increase the equipment count and more space is needed. By higher LNG rate the size
of each unit will also increase. A higher feed gas rate results in a higher refrigerant rate, and hence an
increase in unit size. The heat exchanger will be particularly influenced by higher flow rates. Later it
will be illustrated how each improvement affects the process with a 25 % higher LNG production. The
2DLE case was extended with liquid expander, two stage compression and EFG. The compressors in
the 2DLE case had polytropic efficiencies decided by vendor curves. They all had efficiencies around
80%, so in the simulations with higher LNG rate, all compressors were defined with polytropic
efficiencies at 82%. To meet the LNG specifications, higher heating value was hold constant at 10,95
kWh/m3 (1058 BTU/scf) for both cases. To accomplish a constant higher heating value the
refrigeration duty was varied to keep the nitrogen content below 1 mole%. In the two following cases
the refrigeration duty for the methane cycle was held constant and in the nitrogen cycle the flow rate
was increased.
An increase in LNG production demands a higher duty of the heat exchanger. In the 2DLE the UA
value is constant at 23 860 MJ/°C*h so a higher duty results in a larger ∆T in the heat exchanger. If it
is desired to keep the power consumption down and without additional equipment it is necessary to
increase the heat exchanger size. A higher UA value will allow reduced ∆T for the same duty, thus
reduced exergy losses and reduced power consumption.
Figure 5.3 25% higher LNG production with increase of UA-value
The results represented in Figure 5.3 are based on simulations of the 2DLE with 25% higher LNG
production. The process demanded 0,5502 kWh/kgLNG with the original UA value at 23 860 MJ/°C*h.
Illustrated is the power savings with an increase of the UA value.
41
Table 5.3 Individual improvements in efficiency with new a unit or change in the design (EFG)
In Table 5.3 each improvement is represented alone with 25% increase LNG production. The UA
value is constant at 23 860 MJ/°C*h and only the nitrogen flow rate is adjusted to meet the LNG
specifications. The improvement with a liquid expander stands out as the most efficient solution.
In order to see how the improvements together affected the process they were all simulated in two
cases, with constant UA value and constant LMTD. It was chosen to see how the efficiency of the heat
exchanger affected an extended process with all three improvements. The two cases were based on the
heat exchanger specifications from the 2DLE case with emphasis on its UA value and LMTD. Both
extended processes had the same LNG production and specifications.
Figure 5.4 The extended NicheLNG process
Figure 5.4 illustrates the extended NicheLNG process simulated in HYSYS. Black represent the
natural gas path, green the nitrogen path and blue the EFG. In HYSYS, the process was simulated with
the heat exchanger divided in four parts, HX-1, HX-2, HX-3 and HX-4. Each HX has its own UA
value and LMTD.
42
Table 5.4 Results of 25 % increase in LNG production
Table 5.4 represents four simulated cases. Feed gas enters with 30 bar and 44°C. In all cases the
compressors and expanders have polytropic efficiency of 82% and adiabatic efficiency of 87%
respectively. The 2DLE original is without increase in LNG production and changes in specifications.
In the 2DLE the capacity is increased with 25% and with the same design as in 2DLE original. For
both of the Extended 2DLE the design is illustrated in Figure 5.4. Both cases are either based on the
same UA value or LMTD as in the 2DLE original. Equal LMTD, based on the 2DLE original, was
chosen to see how the UA value affected the efficiency of the process.
43
Figure 5.5 Temperature-enthalpy diagram of the 2DLE with 25% increased capacity
Figure 5.6 Temperature-enthalpy diagram of the Extended 2DLE with constant LMTD and 25% increased capacity
Figure 5.5 and 5.6 illustrate the composite curves for the 2DLE with 25% increased capacity and the
Extended 2DLE with constant LMTD also with 25% increased capacity. From the figures it can be
observed that the Extended 2DLE with constant LMTD demands less duty and the composite curves
are closer in the temperature region below -70°C.
44
5.5
Discussion on increased capacity
Three solutions to make the liquefaction process more efficient have been described. A more efficient
process, and especially with higher production rate, can defend higher investment costs.
After LPG extraction the feed gas entering the liquefaction process has too high nitrogen content to
meet the LNG specifications. Some end flash gas has to be produced and it has a relative high
refrigerant quality due to its low temperature. The refrigerant quality of a gas depends on its capacity
to attract heat and the efficiency of the liquefaction process. With less efficient liquefaction process
any produced EFG will be more valuable. In the 2DLE case with normal LNG production an EFG
utilization would affect the process with 3,3% lower power consumption. Destination of the FPSO is
not decided and the nitrogen content in the feed may therefore vary. The nitrogen content has an
influence on the process and how utilization of EFG will affect power consumption.
The natural gas stream leaving the cold box is already in liquid phase, and still under pressure. By
expansion in a turbine, instead of a valve, power can be recovered. In this case the power recovery is
very small (394 kW), but more important is the temperature drop. When fluid does work by
expansion, at very low temperatures, heat rejection is more valuable. With a liquid expander, the
temperature of LNG leaving the heat exchanger could be increased by 3,4°C. This reduced the
refrigeration load and resulted in a 4,3% lower power consumption.
Feed gas and the methane cycle are compressed by one compressor. This compressor is responsible for
almost half of the total power consumption. With an additional compressor so the compression is done
over two stages with an intercooler, the total power consumption is reduced by 3,2%.
When LNG production was increased by 25% the liquid expander did stand out as the improvement
with highest contribution to efficiency. The liquid expander alone reduced the power consumption
with 5%. The efficiency of the liquefaction part was found to be 26,6% in Chapter 4.4.1. Due to its
low efficiency an improvement that contributes to lower duty of the liquefaction part will have an
increasing influence with higher LNG production. In contrast is the improvement with interstage
compression. Compressors have a relative high efficiency. When LNG production increases the power
consumption of the liquefaction part will grow more rapidly than for the compression. This explains
the decrease from 3,2% to 2,6% when LNG production is increased by 25%. The improvement on
utilizing the EFG production is not influenced by the higher LNG production. With higher LNG
production the EFG production increases with the same rate and the 3,3-3,4% reduction will stay
constant.
An alternative or another contribution to higher efficiency is to operate with a more efficient heat
exchanger. A heat exchanger with higher UA value reduced the power consumption by an average of
0,16% per percent of increased UA value.
The three improvements together were simulated with higher LNG production. Even though
production was increased by 25% it was more efficient than at normal production rate. It applies for
both simulated cases. The two extended NicheLNG processes with 25% higher LNG production will
also affect the heat exchanger. An equally efficient heat exchanger as in 2DLE results in higher UA
value and thereby more necessary space. Depending on available space, a compromise on efficiency
and units or size will decide the final design.
45
46
6
Conclusions and further work
6.1
Conclusions
The NicheLNG process, the chosen liquefaction process for the HLNG FPSO-1, has been described
and evaluated with respect to energy efficiency. The improvement potentials and energy savings have
been presented by thermodynamic analysis and simulations in HYSYS. A comparison with a
promising alternative process has also been presented. Possibilities to expand the NicheLNG process
is considered with increased LNG capacity.
In the evaluation of suitable natural gas liquefaction processes for offshore applications, the expander
processes did stand out as the most promising when emphasis is on; compactness, safety, operation
and equipment count. The most proposed expander process, a dual nitrogen process, was compared
with the NicheLNG process with emphasis on power consumption and energy efficiency. With equal
conditions and LNG production, the NicheLNG process with an exergy efficiency of 31,2% had 10%
lower work consumption. Natural gas as refrigerant has higher cp than nitrogen, resulting in a
significantly lower mass flow rate. Hence, lower flow rate contributes to lower compression work.
The benefit in terms of energy consumption with higher liquefaction pressure depends on the feed gas
pressure and the efficiency of the liquefier. A constant heat rejection temperature and with increasing
liquefaction pressure, the necessary heat to be removed is reduced. The NicheLNG liquefier has an
exergy efficiency of 26,6%. With this efficiency and feed gas pressure in the range of 10 bar to 60 bar,
a higher liquefaction pressure will have a positive influence on the work consumption.
The improvements; utilization of EFG, liquid expander, two stage compression and higher UA-value,
were individually discussed and evaluated with LNG production as normal and 25% higher capacity.
The utilization of End Flash Gas and a new compressor reduced the work consumption by 3,4% and
2,6%, respectively. These improvements reduced the work consumption by a few percent but the
liquid expander, at 25% higher LNG production, stands out alone as the improvement with the highest
contribution. With a liquid expander, the work consumption is reduced by 5%.
The NicheLNG process with 25% higher LNG production demands, in the terms of specific work,
0,5502 kWh/kgLNG. By extending the process with the proposed improvements, the work consumption
is reduced but on the cost of space. The two extended process resulted a specific work of 0,4913
kWh/kgLNG and 0,4791 kWh/kgLNG. They had the same improvements except that the one with the
lowest energy demand differ by its heat exchanger with a 11,2% higher UA value.
47
6.2
Suggestions on further work
In the next phase of the NicheLNG analysis, a more practical view should be evaluated on how the
adjustments and the expansions suggested in this thesis will affect the topside of the FPSO-1. It has to
be taken account of available space and how the changes, on operability, will be influenced. It should
also be evaluated whether the improvements can be justified with respect to investment costs.
A change of the design by closing the open methane cycle, the impact an additional compressor has on
the process should be evaluated. With a closed methane cycle, higher pressures in the cycle can be
chosen. The influence, a higher pressure level has on the unit sizes and the work consumption, should
be investigated.
Since there already are heavy hydrocarbons in liquid phase (LPG) on the topside of the FPSO-1, a
precooler based on propane or a mixture of heavy hydrocarbons can be justified. The benefits, in terms
of reduced work consumption, of a precooler in front of the NicheLNG process must also be evaluated
with respect to investment costs. On the other hand, if it is decided to operate with liquid refrigerants,
a replace of the NicheLNG process with the Single Mixed Refrigerant process could be a promising
alternative.
48
REFERENCES
[1] Moran & Shapiro. “Fundamentals of Engineering Thermodynamics”, 5th edition, 2006, J. Wiley &
Sons Ltd, ISBN 0-470-03037-2.
[2] Kotas T.J. “The exergy method of thermal plant analysis”, Reprint ed. 1995, Fla.: Krieger Pub.,
ISBN 0-89464-941-8.
[3] Venkatarathnam G., “Cryogenic Mixed Refrigerant Processes”, 2008 Springer Science+Business
Media, ISBN 978-0-387-78513-4.
[4] Barclay M. & Denton N. “Selecting offshore LNG processes”, Article presented in LNG Journal
October 2005.
[5] “HLNG FPSO – NTNU.ppt”, Project presentation of Höegh LNG FPSO.
[6] Çengel Y., “Heat and Mass Transfer: A Practical Approach”, 3th edition, 2006, McGraw-Hill,
ISBN-13 978-007-125739-8.
[7] Remeljej C.W. and Hoadley A.F.A.. “An exergy analysis of small-scale liquefied natural gas
(LNG) liquefaction process”, Energy Journal, Volume 31, Issure 12, September 2006 Pages 20052019.
[8] Van de Graaf J.M., Pek B. ”Large-capacity LNG Trains – The Shell Parallel Mixed Refrigerant
Process” Business Briefing: LNG Review 2005.
[9] Nils Jakob Hasle. “HÖEGH LNG FPSO PROJECT – Presentation”, Höegh LNG INTSOK, 4th
March 2009.
[10] Barclay M. and Yang C. “Offshore LNG: The Perfect Starting Point for the 2-Phase Expander?”,
Presentation at the 2006 Offshore Technology Conference 1.-4. May.
[11] Jostein Pettersen. “Natural gas liquefaction process fundamentals”, European Cryogenic Course
– Trondheim June 2009
[12] Hans Quack. “Change of State” - Presentation, European Cryogenic Course - Dresden 2009.
[13] A. Aspelund, D. Berstad, T. Gundersen. ”An Extended Pinch Analysis and Design procedure
utilizing pressure based exergy for subambient cooling”, Applied Thermal Engineering 27 (2007)
2633–2649.
[14] LNG specifications. http://www.nordiclng.no/index.cfm?id=167576 , visited 10.01.2010
[15] DMR for FPSO. http://www.airproducts.com/LNG/ProductsandServices/FloatingLNGPlant.htm ,
visited 19.12.2009.
[16] Nibbelke R.et al. “Double mixed refrigerant LNG process provides viable alternative for tropical
conditions” Based on presentation to 82nd Annual GPA Convention, Dallas, March 11-13, 2002.
[17] Wijngarden W. Leo J. “Offshore Niche LNG production – Unlocking Stranded Gas” Presentation
from KIVI NIVIRA.
[18] A.R. Jha. “Cryogenic Technology and Applications” 2006, Elsevier Inc., ISBN 13: 978-075067887-9.
49
Appendix A
Exergy change and liquefaction work calculated with data from HYSYS simulations.
50
51
52
53
Appendix B
From Aspentech HYSYS Support was ‘Weighted Model’ chosen as heat exchanger parameter. Bellow
is an explanation of the difference between the various heat exchanger models, given by Aspentech.
54
Appendix C
Calculations of the methane refrigeration cycle with different inlet expander temperatures.
55
Appendix D
Feed gas compressed to a liquefaction pressure. Compressor has a polytropic efficiency of 82% and
the liquefiers efficiencies varies.
56
Appendix E-1
Feed gas compressed to a liquefaction pressure. Compressor has a polytropic efficiency of 82%.
Exergy values are exergy change from compressed natural gas to stream out of cold box (Liquefier
100% efficiency).
57
Appendix E-2
Compressions for all calculations are done with a polytropic efficiency of 82%. Feed is pressure
entering compressor after pretreatment and Liq. is outlet compressor pressure/liquefaction pressure.
58
Total work consumption of compression and a liquefier with an efficiency of 26,6%. Feed is pressure
entering compressor after pretreatment and Liq. is outlet compressor pressure/liquefaction pressure.
Graph for the calculations above:
59
Total work consumption of compression and a liquefier with an efficiency of 30%. Feed is pressure
entering compressor after pretreatment and Liq. is outlet compressor pressure/liquefaction pressure.
Graph for the calculations above:
60
Total work consumption of compression and a liquefier with an efficiency of 40%. Feed is pressure
entering compressor after pretreatment and Liq. is outlet compressor pressure/liquefaction pressure.
Graph for the calculations above:
61
Total work consumption of compression and a liquefier with an efficiency of 50%. Feed is pressure
entering compressor after pretreatment and Liq. is outlet compressor pressure/liquefaction pressure.
Graph for the calculations above:
62
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement