United States Patent (10) Patent No.: US 8,381544 B2: Coyle (45) Date of Patent: Feb. 26, 2013

USOO8381544B2
(12) United States Patent (10) Patent No.: US 8,381544 B2

Coyle (45) Date of Patent: Feb. 26, 2013
(54) METHOD FOR LIQUEFACTION OF 3,837,172 A 9, 1974 Markbreiter et al.
NATURAL. GAS 4,009,097 A 2, 1977 Ward
4,479,871 A 10, 1984 Pahade et al.
5,114,451 A 5/1992 Rambo et al.
(75) Inventor: David A. Coyle, Houston, TX (US) 5,755,115 A 5/1998 Manley
5,950,453 A 9, 1999 Bowen et al.
(73) Assignee: Kellogg Brown & Root LLC, Houston, 5,953,935 A 9, 1999 Sorensen
TX (US) 5,956,971 A 9, 1999 Cole et al.
6,016,665 A 1/2000 Cole et al.
c -r 6,023,942 A 2/2000 Thomas et al.
(*) Notice: Subject to any distic the t 6,182,469 B1 2/2001 Campbell et al.
patent is extended or adjusted under 35 6,237,364 B1 5, 2001 Stone
U.S.C. 154(b) by 1254 days. 6,510,706 B2 1/2003 Stone et al.
6,564,579 B1 5/2003 McCartney
(21) Appl. No.: 12/175,559 (Continued)
(22) Filed: Jul.18, 2008 FOREIGN PATENT DOCUMENTS
O O JP 62O72999 A * 4f1987
(65) Prior Publication Data JP 10121071. A * 5, 1998
US 201O/OO11663 A1 Jan. 21, 2010 (Continued)
(51) Int. Cl. OTHER PUBLICATIONS
F25.3/00 (2006.01)
FI 7C 7/02 (2006.01) Crawford, D.B. and Eschenbrenner, G.P. “Heat Transfer Equipment
COIB 3/32 (2006.01) for LNG.” Chemical Engineering Progress, Sep. 1972.
(52) U.S. Cl. .............. 62/618; 62/50.1; 62/50.2: 62/620; Continued
48/127.1; 48/127.3; 137/6 (Continued)
(58) Field of Classification Search ................... 62/50.1, Primary Examiner — Frantz Jules
62/50.2, 618, 620; 48/127.1, 127.3 Assistant Examiner — Joseph Trpisovsky
See application file for complete search history. (74) Attorney, Agent, or Firm — KBRIP Legal Dept.
(56) References Cited (57) ABSTRACT
U.S. PATENT DOCUMENTS A method of altering the heating value of a liquefied natural
2,815,650 A 12/1957 McIntire et al. gas by adding higher heating value components is disclosed.
2.952,984 A 9, 1960 Marshall et al. A portion of the liquefied natural gas is used to cool the higher
3,405,530 A 10, 1968 Denahan et al. heating value component stream prior to combining the
3,407,052 A * 10, 1968 Huntress et al. ............. 48,127.3 higher heating value components with the liquefied natural
3.420,068 A 1/1969 Petit et al.
3,446,029 A 5/1969 Grgurich et al. gas to obtain a combined stream having a heating value
3,570,261 A 3, 1971 Schwartzman greater than the liquefied natural gas.
3,656,312 A 4, 1972 Streich
3,760,597 A * 9, 1973 Becker ........................... 62/48.2 22 Claims, 18 Drawing Sheets
900
US 8,381544 B2
Page 2
U.S. PATENT DOCUMENTS Lynn, S., Quinlan, M.P., Velasquez, D., and Leppin, D. "New
6,604,380 B1 8, 2003 Reddick et al. Approach to Sulfur Removal Could Reduce Costs.” GasTIPS, Winter
2005, 0126220 A1* 6/2005 Ward .............................. 62/625 2002.
2005/0204625 A1* 9, 2005 Briscoe et al. 48,127.3 Jones, R. Sawchuk. J., and Davis, K. “Big Green Train.” AIChE
2006/0260468 A1* 11/2006 Amin .............................. 95/153 Spring Meeting, 2003.
2008/0202161 A1* 8/2008 Vazquez-Esparragoza Kotzot, Heinz J. “Overview of the LNG Industry Seminar: Gas Treat
et al. ............................... 62,620 ment, Liquefaction, and Storage. GPA Convention, San Antonio,
2009/0025422 A1 1/2009 Sicinski et al. ................. 62,613 Mar. 2003.
FOREIGN PATENT DOCUMENTS Jones, R., Sawchuk. J. Durr, C. and Davis, K. “BP's Big Green Train:
JP 10185886 A * 7, 1998 Benchmarking Next Generation LNG Plant Designs.” LNG 14.
JP 2001208297 A * 8, 2001 Doha, Qatar, Mar. 2004.
Durr, C. and de la Vega, F. “Designing Safety into LNG Export?
OTHER PUBLICATIONS Import Plants.” LNG 14, Doha, Qatar, Mar. 2004.
Shah, J.T. and de la Vega, F. "Advisory System for LNG Storage Jamieson, A. and Durr, C. “Nigeria LNG Overcoming Technical
Facilities Uses KBS Technology.” Expert Systems Applications, Oct. and Commercial Challenges to Become One of the World's Major
1991.
Exporters of LNG.” LNG 14, Doha, Qatar, Mar. 2004.
Songhurst, B. “LNGUpdate.” GPA Europe Conference, Dublin, May
Collins, C. “Developments in LNG Trade & Technology.” GPA Euro 19-21, 2004.
pean Chapter LNG Meeting, London, Nov. 28, 1991. Durr, C., Coyle, D., Hill, D., and Smith, Sharon, “LNG Technology
Quinlan, M.P. and Echterhoff, L.W. “Technical and Economic Com for the Commercially Minded.” GasTech 2005.
parison of LO-CAT II with Other Iron-based Liquid Redox Pro Shah, P. and Durr, C. “Monetizing Stranded Gas.” SPE Handbook
cesses.” GRI Liquid Redox Sulfur Recovery Conference, Austin, Chapter, 2005.
Oct. 4-6, 1992.
McKee, R.L., Echterhoff, L.W., and Hertz, D.W. “Control of Aro Cho, J., de la Vega, F. Kotzot, H., and Durr, C. “Innovative Gas
matics Emission for Glycol Dehydration Units: A Look at Novel Processing with Various LNG Sources.” LNG Journal, Jan./Feb.
Approaches.” Summer National AIChE Meeting, Seattle, Aug. 2005.
15-18, 1993. Townsend, B. and Khaligh, B. “Opportunities for Elegant Power
de la Vega, F. Durr, C., Hill, D.K., and Collins, C. “Plant Reliability Systems Design in LNG.' SMi LNG Conference, London, Feb.
Analysis in LNG Plants.” LNG 11 Conference, Birmingham, UK, 21-22, 2005.
1995. Durr, C., Caswell, C. Echterhoff, L., Hill, D.K., and Coyle, D.A.
Collins, C. “LNG Technologies, Challenges, Regulations, and “LNG Technology is on the Move: The Only Thing Constant is
Safety.” IBD Conference, London, Dec. 1997. Change . . .” GasTech, Bilbao, Spain, Mar. 2005.
Rios, J., Coyle, D.A. Durr, C.A., and Frankie, B. “Removal of Trace Coyle, D.A. and Patel, V. "Processes and Pump Services in the LNG
Mercury Contaminants from Gas and Liquid Streams in the LNG and Industry,” 22nd International Pump Users Symposium, Houston,
Gas Processing Industry.” GPA 77th Annual Convention, Mar. 16-18, Feb. 28-Mar. 3, 2005.
1998. Attaway, D.A., Huang, S. Kotzot, H., and Durr, C.A. “Optimal
Collins, C. and Durr, C.A. “Cost Reduction for LNG Export & Import Process Location for NGL Recovery in LNG Plant.” AIChE Spring
Facilities.” Prepared for the SMi Conference Understanding and National Meeting, 2005.
Improving the Economics of Liquefied Natural Gas, London, Jan. Coyle, D. Durr, C., and Shah, P. “LNG: A Proven Stranded Gas
25-26, 1999. Monetization Option.” SPE Annual Technical Conference and Exhi
Mayer, M., Robertson, A. Sheffield. Courtay, R., and Harr, R. “Safe bition, Denver, Oct. 5-8, 2005.
Production of LNG on an FPSO, GasTech 2000. Coyle, D., de la Vega, F. Durr, C. “Natural Gas Specification Chal
Patel, D., Tarlowski, J., and Collins, C. “Developments in LNG lenges in the LNG Industry.” LNG 15 Conference, Barcelona, Spain,
Storage.” SMI Conference, Feb. 2000. Apr. 24-27, 2007.
Sheffield, J.A. and Mayer, M. “The Challenges of Floating LNG
Facilities.” GPA Spring Meeting, Norwich, 2001. * cited by examiner
U.S. Patent Feb. 26, 2013 Sheet 1 of 18 US 8,381544 B2
C
tf
N
ven
d)
a ym
S.N C
N
t en
Se N
N ea p S
ra N
is . . . . . . . . g . . . . . . . g g g g g (
C
er
gas \
& gas
sa
\| - a
/
S
Y
U.S. Patent
do 'e.pneuedue
/
3.
3 TWA NIWGH 3 ?h
706
f / 700
70
H
31
I
Wad
C
ve do
es
o
20
s c co
co
S.
CO
A. y
N
N a
b)
u
co
/
ra
f
da & o
Wo
C C.
o
S.
e
/
Nd
N ya
b)
r
\-
e o
L
cC C d
Ca o
S
S
US 8,381544 B2
1. 2
METHOD FOR LIQUEFACTION OF FIG. 12 depicts two illustrated examples of full contain
NATURAL. GAS ment LNG storage tanks.
FIG. 13 depicts two illustrated examples of membrane
BACKGROUND LNG storage tanks.
FIG. 14 depicts two illustrated examples of cryogenic con
1. Field crete LNG storage tanks.
The present embodiments generally relate to liquefied FIG. 15 depicts two illustrated examples of spherical LNG
hydrocarbon fluids, and to methods and apparatus for pro storage tanks.
cessing such fluids. Natural gas is an important energy source FIG. 16 depicts a schematic illustration of a conventional
that is obtained from Subterranean wells; however, natural gas tap into a vessel.
is sometimes impractical or impossible to transport natural FIG. 17 depicts a schematic illustration of an alternate
gas by pipeline from the wells where it is produced to the sites thermosyphon tap arrangement.
where it is needed, due to excessive distance or geographical FIG. 18 depicts a schematic illustration of a hybrid amine/
barriers such as oceans. In Such situations, liquefaction of membrane CO2 extraction system.
natural gas offers an alternative way of transporting natural
15 FIG. 19 depicts a schematic illustration of an embodiment
gaS.
of the present invention wherein a cooled stream of LPG is
added to an LNG stream to increase the heating value.
2. Description of the Related Art
Natural gas can be converted to liquefied natural gas (LNG) DETAILED DESCRIPTION
by cooling it to about -161° C., depending on its exact com
position, which reduces its volume to about /600th of its A detailed description will now be provided. Each of the
original value. This reduction in Volume can make transpor appended claims defines a separate invention, which for
tation more economical. The liquefied natural gas (LNG) can infringement purposes is recognized as including equivalents
be transferred to a cryogenic storage tank located on an to the various elements or limitations specified in the claims.
ocean-going ship. Once the ship arrives at its destination, the 25 Depending on the context, all references below to the “inven
LNG can be offloaded to a regasification facility, in which it tion may in some cases refer to certain specific embodiments
is converted back into gas by heating it. Once the LNG has only. In other cases it will be recognized that references to the
been regasified, the natural gas can be transported by pipeline “invention' will refer to subject matter recited in one or more,
or other means to a location where the natural gas can be used but not necessarily all, of the claims. Each of the inventions
as a fuel or a raw material for manufacturing other chemicals. 30 will now be described in greater detail below, including spe
cific embodiments, versions and examples, but the inventions
BRIEF DESCRIPTION OF THE DRAWINGS are not limited to these embodiments, versions or examples,
which are included to enable a person having ordinary skill in
So that the manner in which the above recited features of the art to make and use the inventions, when the information
the present invention can be understood in detail, a more 35 in this patent is combined with available information and
particular description of the invention, briefly summarized technology.
above, may be had by reference to embodiments, some of One embodiment of the present invention is a method of
which are illustrated in the appended drawings. It is to be altering the heating value of a liquefied natural gas stream that
noted, however, that the appended drawings illustrate only includes providing a first liquefied hydrocarbon composed of
typical embodiments of this invention and are therefore not to 40 at least 90 wt % methane, having a temperature of about
be considered limiting of its scope, for the invention may -150° C. or less and having a first heating value, in a first
admit to other equally effective embodiments. storage vessel and providing a second liquefied hydrocarbon
FIG. 1 depicts a block flow diagram on one embodiment of composed primarily of one or more of ethane, propane, and
an LNG liquefaction system. butane, or mixtures thereof. A first stream of the first liquefied
FIG. 2 depicts a schematic illustration of a dehydration 45 hydrocarbon is provided from the first storage vessel, from
unit. which a portion is compressed to form a second stream having
FIG. 3 depicts a schematic illustration of a NGL/LPG a pressure higher than the first stream. Passing the second
fractionation unit. stream through the cold side of a first heat exchanger to obtain
FIG. 4a depicts a graph showing the condensation curve of a third stream of first liquefied hydrocarbon with a higher
a mixed refrigerant system. 50 temperature than the second stream and decompressing the
FIG.4b depicts a graph showing the condensation curve of third stream in a first liquid expander to obtain a fourth
a nine level cascade pure refrigerant system. stream. The first liquefied hydrocarbon of the fourth stream is
FIG. 5 depicts a schematic illustration of a conventional returned back to the first liquefied hydrocarbon storage ves
refrigeration system for an LNG facility. sel. A fifth stream of the second liquefied hydrocarbon is
FIG. 6 depicts a schematic illustration of an alternate 55 provided that is compressed to obtain a sixth stream which is
refrigeration system for an LNG facility. passed through the warm side of the first heat exchanger to
FIG. 7 depicts a schematic illustration of one embodiment cool the sixth stream and obtain a seventh stream of second
of an endflash unit for nitrogen rejection. liquefied hydrocarbon with a lower temperature than the sixth
FIG. 8 is a graph illustrating the changing heating value of stream. A first portion of the seventh stream is combined with
fuel gas in a step change and a ramping change. 60 the first stream to obtain an eighth stream having a second
FIG.9 depicts a schematic illustration of one embodiment heating value greater than the first heating value of the first
of a mixing vessel to control heating value changes in fuel liquefied hydrocarbon.
gaS. The first liquid expander can provide static expansion to
FIG. 10 depicts two illustrated examples of single contain the third stream to obtain the fourth stream. The method can
ment LNG storage tanks. 65 also include diverting a second portion of the seventh stream
FIG. 11 depicts two illustrated examples of double con and decompressing that stream in a second liquid expander to
tainment LNG storage tanks. form a ninth stream and injecting the ninth stream into the
US 8,381544 B2
3 4
fifth stream. The second portion of the seventh stream can liquefaction plant which cools the natural gas to about -161°
range from 0% to 100% of the seventh stream. The second C., and reduces its Volume to about/600th of its original value.
liquid expander can provide static expansion to a portion of Liquefaction plants are generally built as either peak-shaving
the seventh stream to obtain the ninth stream. or as base-load plants, depending on their function in the
The flow of the seventh stream into the first stream can be overall natural gas flow. Peak-shaving facilities are relatively
regulated to control the heating value of the eighth stream. Small (in the hundred thousand tons per year range) to provide
The temperature of the seventh stream in some embodiments a buffer between a more constant Supply and a highly fluctu
is no more than 50° C. warmer than the temperature of the first ating demand. These facilities typically liquefy natural gas
stream. The temperature of the seventh stream in some during periods of low demand and vaporize natural gas at
embodiments is no more than 25 C warmer than the tempera 10 times of peak demand. Base-load plants liquefy several mil
ture of the first stream. The high heating value of the eighth lion tons of natural gas per year for ship transportation to the
stream in embodiments can be 1050 Btu/SCF (39.1 MJ/Sm3) consumer. A few small plants also exist for the purpose of
or greater.
The pressure of the third stream can be sufficient to keep delivering LNG locally by road or rail tanker when alternate
the third stream in a liquid State and in some embodiments is transport is deemed more economical than pipeline delivery.
no warmer than -100° C. The temperature of the seventh 15 Although the construction of a liquefaction plant is typi
stream can be cold enough to prevent cavitations as the sev cally the single largest cost element, large investments are
enth stream is combined with the first stream. also needed throughout the gas Supply chain in areas Such as
The temperature approach on the first heat exchanger exploration, construction of specialized ships for LNG trans
between the third stream and the sixth stream can be less than portation, and in receiving terminals for unloading of LNG
20°C. The temperature approach on the first heat exchanger from the ships and regasification of the LNG to natural gas in
between the second stream and the seventh stream can be less vapor form. The main technical challenge is to minimize the
than 20°C. The flow of the ninth stream can be regulated to costs of a Supply chain while maintaining the high level of
maintain a temperature approach on the first heat exchanger safety, availability, and longevity. The LNG supply chain
between the third stream and the sixth stream of less than 20° typically includes the following processes: natural gas pro
C. The flow of the ninth stream can be regulated to maintain 25 duction and transportation to the liquefaction facility: gas
a temperature approach on the first heat exchanger between treatment; liquefaction of the gas to LNG: LNG storage;
the second stream and the seventh stream at no more than 20° loading of LNG onto transport vessels; transportation;
C. The method can further include loading the eighth stream unloading of LNG at a receiving terminal; LNG storage;
onto an LNG transport vessel. The temperature of the fifth regasification of the LNG and delivery to a gas pipeline; or an
stream can be 0 C or less or can be -30°C. or less. The first 30 onsite electrical power generation facility.
heat exchanger can be one or more plate-fin type exchangers. Feed Gas
An alternate embodiment of the present invention is a Natural gas can be produced from a field of subterranean
method of modifying the heating value of a liquefied natural wells. Raw natural gas contains primarily methane. Raw
gas stream that includes providing a first liquefied hydrocar natural gas also can contain Smaller amounts of ethane, pro
bon (LNG) composed of at least 90 wt % methane, and having 35 pane, n-butane, isobutane, and heavier hydrocarbons, as well
a first heating value; providing a second liquefied hydrocar as water, nitrogen, helium, mercury, and acid gases such as
bon (LPG) composed primarily of one or more of ethane, carbon dioxide, hydrogen Sulfide, and mercaptains. Raw natu
propane, and butane, or mixtures thereof, passing a first ral gas can be classified as either associated gas (which is
stream of the LNG from the first storage vessel, through a first produced from reservoirs that also contain Substantial
heat exchanger to provide cooling energy; passing a second 40 amounts of oil) or non-associated gas (which is produced
stream of the LPG through the first heat exchanger to obtain from reservoirs that generally contain little or no oil).
a third stream of LPG with reduced temperature; providing a Components that are commonly removed from the feed gas
fourth stream of LNG; blending a first portion of the LPG of include: carbon dioxide, hydrogen sulfide, Sulfur compounds,
the third stream with the LNG of the fourth stream to obtain aromatics, water, mercury, nitrogen and heavier hydrocar
a fifth stream of LNG containing increased content of LPG 45 bons. A mixture of ethane, propane, n-butane, isobutane, and
and thus having a heating value greater than the first heating in some cases heavier hydrocarbons can be separated from
value of the first liquefied hydrocarbon. natural gas and recovered as a product referred to as natural
The method can further include returning the warmed first gas liquids (NGL). A mixture primarily propane and butanes
stream exiting the first heat exchanger to the first storage can also be recovered as a product referred to as liquefied
vessel. The first heat exchanger can comprise one or more 50 petroleum gas (LPG). Natural gas that has a relatively high
plate-fin type exchangers. The temperature of the third stream NGL and/or LPG content is sometimes referred to as "rich'
exiting the first heat exchanger in some embodiments can be gas and natural gas that contains a relatively low NGL and/or
no more than 50° C. warmer than the temperature of the first LPG concentration is sometimes referred to as “lean' gas.
stream entering the first heat exchanger, or no more than 25° In some embodiments, some processing of the produced
C. warmer than the temperature of the first stream entering the 55 gas can be performed near the well; for example, in order to
first heat exchanger or no more than 10°C. warmer than the prevent or minimize dropout of hydrocarbon liquids, forma
temperature of the first stream entering the first heat tion of hydrates, and pipeline corrosion during transportation
exchanger. The method can further include blending a second from the well to a plant at which more extensive processing
portion of the third stream exiting the first heat exchanger into can take place. After any processing that takes place in or near
the second stream prior to entering the first heat exchanger. 60 the well, the natural gas can be transported to a liquefaction
The method can further include loading the fifth stream onto plant, for example by pipeline.
an LNG transport vessel. With reference to the figures, FIG. 1 depicts a block flow
diagram of one embodiment of an LNG liquefaction system.
Illustrative Embodiments The depicted generalized LNG liquefaction plant block flow
65 diagram illustrates the major components of an overall LNG
Liquefaction of natural gas to produce the product lique liquefaction facility 10 Such as a gas treating section 100, a
fied natural gas (LNG) can be conducted in a facility called a liquefaction/refrigeration section 200, and an LNG send out
US 8,381544 B2
5 6
and storage section 300. A gas treating section 100 can com Some processes used for removal of carbon dioxide and
prise gas reception facilities 110, acid gas removal unit 120, sulfur compounds include the use of MEA (monoethanola
and dehydration/mercury removal unit 130. The liquefaction mine), MDEA (methyl diethanol amine), and proprietary
section 200 can comprise an initial cooling/condensing unit compounds such as a-MDEA (BASF), Sulfinol (Shell) and
210 to remove heavier hydrocarbons, liquid removal with the Benfield Hi Pure process and Mol Sieve Benfield (MSB)
fractionation 220, liquefaction 230, refrigeration system 240, process (UOP). The latter process also includes dehydration
and endflash/nitrogen rejection unit 250. An LNG send out of the feed gas by co-adsorption of water and residual carbon
and storage section 300 can comprise storage for the LNG dioxide leaving the hot carbonate absorption section. The use
310, LNG/LPG 320, and heavier hydrocarbon liquids 330 of membrane/amine and cryogenic fractionation schemes
that are sometimes referred to as gasoline. The acid gas 10 have also been proposed for high CO feed gas. In certain
removal unit 120 can remove hydrogen sulfide, carbon diox installations two absorbers have been provided in each pro
ide, and other impurities via line 122. The dehydration/mer cess train. The use of packing, including structured packing,
cury removal unit 130 can remove water and mercury as Sometimes enables the use of a single absorber vessel for a
single process train with a high feed gas rate. When consid
illustrated via line 132. The endflash/nitrogen rejection unit 15 ering these and other processes, particular attention must be
250 can remove nitrogen as illustrated via line 252. In some given to the feed gas composition, product specifications and
facilities, a helium-rich stream also is produced for further the distribution of impurities between the various streams
processing in a helium plant. used for make up to refrigerant circuits.
Gas Treating Membrane, Amine and Hybrid Systems: Membrane sys
Carbon dioxide, water, aromatics, and heavier hydrocar tems have shown to be an economically and operationally
bons can freeze on the heat transfer Surfaces of heat exchang viable alternative for gas processing. An economic compari
ers at cryogenic temperatures. Mercury can cause degrada Son of membrane, amine, and membrane/amine hybrid pro
tion of aluminum, which is used in Some low temperature heat cessing schemes can be conducted to determine when instal
exchangers. Excessive concentrations of Sulfur can also be lation of a hybrid system may be advantageous for CO
undesirable in natural gas and their removal may be dictated 25 removal associated with LNG production.
by the end-user gas specification. Therefore, the gas treating Amine units increase their energy usage when the CO
section of a liquefaction plant can include apparatus for concentration increases. Membranes, on the other hand, use
removing at least a portion of these substances. In one partial pressure as the driving force and are most effective at
embodiment, a gas treating section can comprise an inlet/pre high concentrations of CO. In a hybrid system the membrane
treatment section, an acid gas removal section, a dehydration 30 can remove the bulk of the CO, with a relatively small area
section, a mercury removal section, and a particle filtration and the amine is used for the final cleanup to achieve the
section. required CO specification. The advantage of this is that bulk
Gas Reception Facilities: The inlet/pre-treatment section removal with membranes may reduce capital and operating
can remove solid and liquid phase materials and impurities costs when compared to conventional processes. The savings
from the natural gas feed stream. These facilities can separate 35 between membrane and amine increases as the CO content
out heavy liquid hydrocarbons that have condensed from the increases. Increasing CO partial pressure favors membrane
natural gas. This liquid can be primarily a Cs and higher systems, whereas, amine Solution loading is limited by cor
condensate, and can be sent either to a NGL fractionation unit rosivity considerations. As the feed gas price increases, the
or directly to NGL storage tanks. A condensate stabilization hybrid systems range of operability increases. This trend
facility can remove any associated sour gas from the conden 40 occurs, again, because the membrane system's economics are
sate and can compress it back into the natural gas stream. The governed by the loss of hydrocarbons. It is important to note
pressure of the feed gas can be reduced at this point if the feed that the hybrid system does not take applicability away from
gas pipeline pressure is higher than the LNG plant operating the membranes but rather increases the range of applicability.
pressure. Alternatively, feed gas arriving at low pressure can FIG. 18 depicts a schematic illustration of a hybrid amine/
be compressed to reach the desired LNG plant operating 45 membrane CO2 extraction system. The hybrid process 800
pressure. includes a feed gas stream 802 that flows through a membrane
Particle Removal: A particle filtration section can reduce 804 that removes acid gas via line 806 and a partially pro
the concentration of particulate matter, for example by filtra cessed gas stream via 808. The partially processed gas stream
tion, in order to prevent plugging of downstream equipment. 808 enters an amine contactor 810 that produces processed
Acid Gas Removal Unit: The acid gas removal Section can 50 gas via line 812. CO lean amine enters the amine contactor
remove carbon dioxide and Sulfur-containing components 810 via line 814 while CO rich amine leaves the amine
Such as hydrogen Sulfide that are present in the natural gas contactor 810 via line 816 and flows to an amine stripper 820
feed. Acid gases can be removed by processes such as treat where CO and possibly other gasses are released by heating
ment with chemical Solvents, physical solvents, adsorption and reducing pressure and leaves the amine stripper 820 via
processes, or physical separation. For example, amine pro 55 line 822.
cesses can be used for acid gas treatment, in which an aqueous Sulfur Extraction: Research into catalysts that can effect
Solution of an alkanolamine is contacted with the gas and the direct oxidation of hydrogen sulfide to sulfur has been
removes at least some of the Sulfur compounds. Off gases conducted, for example by TDA Research. Direct oxidation
from the acid gas treatment unit can be sent to a Claus unit for of hydrogen Sulfide in a single reactor vessel, capable of
sulfur recovery if the recoverable sulfur is sufficiently large. If 60 processing less concentrated Sulfur streams, could be benefi
the quantity of sulfuris Small, the off gas can be incinerated or cial because of its simplicity and small footprint. Modified
the Sulfur in the acid gas stream can be further enriched in an Claus processes have many commercial applications, but
acid gas enrichment process upstream of a Claus unit. Off gas typically require a more concentrated hydrogen Sulfide
also can be further treated with one or more Scavenger pro Stream.
cesses. MDEA, usually applied with a proprietary activator to 65 Dehydration
facilitate gas pickup, tends to be less corrosive and has lower The dehydration section can remove water from the gas to
regeneration heat duties than other amines. prevent hydrate formation and freezing in the cryogenic sec
US 8,381544 B2
7 8
tion of the plant. The dehydration unit sometimes removes Mercury Removal
sulfur compounds also, such as RSH, COS and CS. FIG. 2 Mercury, often present intrace amounts, attacks piping and
depicts a schematic illustration of a dehydration unit. In FIG. equipment made from aluminum and aluminum compounds.
2, the inlet gas 134 can be cooled to a point above the hydrate The mercury removal section can lower the mercury concen
formation temperature using air or water coolers 136 and tration in the natural gas to prevent corrosion of aluminum
refrigerant. Free liquid water that is condensed can be sepa equipment in the liquefaction process. Mercury must be
rated in a separator 140 and removed via line 142, hydrocar removed prior to the feed gas entering the cryogenic sections
bon liquids can be separated and removed via line 144 and the of the LNG plant. The mercury removal vessel is usually
remaining gas can be fed via line 146 to one or more molecu placed after the dehydration unit, but can also be located
lar sieve adsorption unit(s) 150 for removal of water vapor 10 upstream of an acid gas removal unit. Mercury can be
down to the desired maximum concentration. The dried gas removed from the gas by reaction with elemental sulfur to
leaves the molecular sieve adsorption unit(s) 150 via line 152 form a sulfide. The sulfur can be supported on a high surface
and can pass through one or more filters 154, and leaves via area solid carbon bed. This can be a non-regenerative process
line 156 to the liquefaction section. A regeneration heater 160 and the spent carbon bed can be disposed of in a landfill or can
can be used to heat a portion of the dehydrated gas 162 for 15 be returned to a catalyst vendor for reclaiming. Filtering of the
flow through the molecular sieve adsorption unit(s) 150 for gas stream is typically required to prevent any solid particles
regeneration. After the heated gas flows through the molecu from being carried with the gas into the Subsequent sections
lar sieve adsorption unit(s) 150 they exit via line 164 and to a of the liquefaction plant.
regeneration cooler 166, a regeneration gas knockout drum Mercury Forms: Mercury is present in natural gas and
168 and a regeneration gas compressor 170 and are injected natural gas associated condensates, as organometallic and
back into the inlet gas stream. Options for dehydration in inorganic compounds, and in the elemental (metallic) form
various embodiments can include two or three dehydration depending on the origin of the gas. The elemental form can be
beds, high or low regeneration pressure (high pressure regen found in either the vapor or liquid phase. The organometallic
eration gas requires only little re-compression if returned to (typically dimethyl mercury, methylethyl mercury, or diethyl
the acid gas removal section). Other options include the 25 mercury) and inorganic (such as HgCl2) compounds drop
regeneration gas source; the Source to use for regeneration into the liquid phase in any fractionation of the natural gas
heat (for example, heat can be provided from various heat streams. Vapor phase elemental mercury is a primary culprit
Sources available in the plant); and regeneration gas to fuel in corrosion of aluminum exchangers inside cryogenic cold
instead of its reinjection into an inlet gas stream. boxes. Elemental mercury that leaves the plant in the liquid
Drying with fixed-bed desiccants: When activated desic 30 phase natural gas condensate streams is a primary source of
cant is brought into contact with a gas of high humidity, the corrosion in aluminum equipment in olefin plants that crack
pressure of the water vapor in the adsorbent tends to reach NGL recovered from the natural gas plants. Mercury also
equilibrium with the pressure of the water vapor in the sur poisons the selective hydrogenation catalysts in olefin plants,
rounding gas. Hence the moisture content of the gas is and can pose inhalation hazards to workers. Organometallic
decreased. Physical adsorption is a continuous process. 35 and inorganic mercury usually end up in the condensate
Adsorbed molecules break away from the adsorbent when stream from the natural gas plant. These compounds are
they acquire sufficient energy. When the rate at which water important environmental toxins that are easily absorbed and
molecules leave the adsorbent is equal to the rate at which accumulated by biological organisms. The presence of these
they are being adsorbed on the Surface, equilibrium is estab compounds in natural gas condensate streams leads to waste
lished and no further adsorption takes place. The amount of 40 disposal problems and safety hazards to workers.
water adsorbed at equilibrium increases with higher partial Mercury Detection: Problems with mercury detection
pressure of the water in the gas, and decreases with higher mean that operators will often have no indication of impend
temperatures. The desiccant is said to be saturated (for a given ing trouble until failure of an equipment item due to mercury
set of conditions) when equilibrium is attained, and the induced corrosion. Since mercury can be present in very low
amount of water adsorbed at this point is known as the equi 45 levels in natural gas streams it is difficult to determining
librium capacity or static capacity. In an embodiment of a which streams are contaminated, and the degree of contami
dehydration unit the water vapor of the processed gas is nation. Detection methods continue to improve, however. A
reduced to below 0.1 ppmv. number of available analyzers now claim capability at the
At the start of operation with a freshly reactivated desic parts per trillion by volume (pptv) level. These methods can
cant, the moisture or other adsorbate is removed from the 50 comprise passing a sample stream through a mercury trap
flowing gas stream at the inlet section of the bed. The distance (dosimeter) over a long collection period, and then desorbing
required for this removal is the depth of bed known as the the mercury from the trap as a concentrated pulse into a
adsorption Zone. This Zone moves through the bed at a uni detector. Primary methods for elemental mercury detection in
form rate depending on the operating conditions, and the the gas include gold filament analyzers, cold vapor atomic
effluent moisture from the bed will be at a low value until the 55 fluorescence (CVAF), peroxide scrubbing, and ICP/mass
adsorption Zone reaches the exit end of the bed. A small spec. CVAF can be a preferred method for laboratory work, as
amount of additional adsorption occurs at the inlet section; CVAF is very sensitive. It has been reported that accurate and
but the rate of adsorption is relatively low after the adsorption consistent results measuring mercury levels in the feed and
Zone has passed, and there is usually insufficient time for treated gas can be obtained using the Jerome Model 431
equilibrium to be established between gas and desiccant. 60 mercury analyzer to less than 0.001 ug/Nm3 using the gold
When the leading edge of the adsorption Zone reaches the exit wire trap, provided that the sample lines are kept ultra-clean
end of the bed, the effluent moisture rises rapidly from then over the required long collection period. Another company
on. Depth of the adsorption Zone and the rate at which the installed a continuous online sampling system that could
adsorption Zone moves through the bed will be determined by accurately measure mercury downto the 0.001 ug/Nm3 level,
a number of variables. These include gas temperature and 65 and found that the Sir Galahad system by PSAnalytical Ltd.
moisture content, desiccant type and particle size, bed depth (atomic fluorescence spectrometry) unit could be adapted. All
and configuration, flow rate and pressure of the gas. of the methods mentioned above are gradually being
US 8,381544 B2
9 10
improved to more accurately detect ppbv levels in the feed effectiveness at bulk removal. However, physical separation
gas, and the pptV required in the treated gas. However they cannot remove the mercury that absorbs into Soil, or mercury
still have some problems, for example, gold filament analyZ chemically bonded to Scavenger catalysts. Biological treat
ers, which are the most frequently used detection instrument, ments use bacteria to concentrate organic mercury com
can have trouble measuring organometallic mercury, are sen pounds. Immobilization technologies reduce the leach ability
sitive to temperature changes, moisture, H2S, and mercaptains of mercury into groundwater from contaminated soils.
in the gas, and cannot measure the gas at operating pressures. Chemical treatments show promise in providing an alterna
For liquid analysis there are two known models: Nippon tive to thermal treatment. These technologies typically use an
Instrument Cop. Model SP-3D (uses an atomic absorption acid to leach the mercury from the contaminated material.
detector), and the PS Analytical Ltd Merlin model (which 10 New approaches suggest using an organic chelating agent
uses an electron fluorescence detector). One company had the from which the mercury can be recovered.
analytical capability to analyze liquid samples including Natural Gas Specification
naphthas and natural gas condensates, and had developed a In some embodiments of the process, the natural gas after
method to differentiate between the various species. Levels as treatment of can have the following maximum concentrations
low as 0.1 ppb wt. can be quantified. The most accurate 15 of the listed components:
measurements should be done at operating conditions and a) hydrogen sulfide 3-4 ppmv
over a prolonged collection period. As a further complication b) total sulfur 30 mg per normal cubic meter
to the situation, mercury levels in natural gas have been c) carbon dioxide 50 ppmv
reported to fluctuate by a factor of five over periods longer d) mercury 0.01 ug per normal cubic meter (0.013 ppb by
than eight hours. In addition, when samples are collected in weight)
the field and brought to the lab, some of the mercury is e) nitrogen 1 mol %
adsorbed on the container walls, resulting in lower readings. f) water vapor 1 ppmv
Scavenging Elemental Mercury from the Vapor Phase: g) benzene 1 ppmv
Scavenging elemental mercury and organometallic com h) ethane 6-8 mol %
pounds from the feedstock(s) of gas plants is a maturing 25 i) propane 3 mol %
technology for mercury removal. Elemental mercury can be j) butane and heavier hydrocarbons 2 mol %
readily trapped by contact with Sulfur based trapping agents. k) pentane and heavier hydrocarbons 1 mol %
The principle commercial trapping agents include Sulfur, The desired maximums for the various components can
metal sulfides, silver and gold. The types of support materials vary, depending for example on the desired heating value of
or carrier agents include activated carbon, alumina and other 30 the gas. In some embodiments, the high heating value of the
Zeolite materials. Operating temperatures range from ambi gas can be from about 1050 to about 1100 BTU/SCF. In some
ent to 100 C, and at pressures up to 100 bar. Elemental embodiments the high heating value can be specified to be
mercury in the gas phase is readily trapped by Sulfur based 1065 BTU/SCF or greater.
trapping materials which fixes the Volatile mercury in the Liquefaction/Refrigeration Section
form of non-volatile mercury sulfide (HgS). Most commonly, 35 Referring back to FIG. 1, in one embodiment, a liquefac
an activated carbon is chemically treated or impregnated with tion/refrigeration section 200 can comprise a feed gas cooling
a mercury-fixing compound such as Sulfur or iodine. The and NGL condensation section 210, a fractionation section
mercury is chemisorbed onto the non-retentive carbon which 220, a liquefaction section 230, a refrigeration section 240,
must be periodically replaced (typically every 3-4 years). and an endflash section 250.
Removing Organometallic Mercury from the Liquid 40 Initial Cooling to Remove Heavier Hydrocarbons: The
Phase: Less work has been done on the problem of removal of feed gas cooling and NGL condensation section 210 can
organometallic mercury from liquid streams than the remov separate ethane and heavier hydrocarbons from the feed gas
ing of elemental mercury in the vapor phase. The current to adjust the heating value and to remove benzene. Initial
leading approaches involve adsorption onto a carbon or chilling can be provided, for example, by the first stage of a
molecular sieve, and the use of ion exchange resins, such as: 45 propane refrigeration cycle (e.g., cascade and propane pre
ALCOA Mersorb LH (Impregnated pelletized activated car cooled mixed refrigerant) or in a warm heat exchanger of a
bon); Calgon Type HGR-IS (Potassium iodide impregnated mixed refrigerant (MR) process. For this initial cooling, kettle
granular activated carbon); Stamicarbon Ion Exchange Pro or core-in-kettle type exchangers can be used if the tempera
cess (Ion exchange resin with thiol groups); ICI Katalco ture remains above the minimum for carbon Steel applica
n/1E32LEsPECTM 1157 (a fixed bed chemical absorbent 50 tions. At lower temperatures, for which special materials can
mixed metal oxides with cementitious binder that can be used be required, plate fin (PF) or spiral wound (SW) aluminum
for removal of elemental mercury in liquid streams such as exchangers can be used.
NGL liquids and LPG); and/or UOP HgSIV (Regenerative Precooling and Liquefaction: One process development
molecular sieve containing silver that can be used for removal involves replacement of propane precooling with a multicom
of elemental mercury in liquid streams such as NGL liquids 55 ponent refrigerant system. This alternative enables precool
and LPG). ing to a lower temperature than that allowed by using atmo
Another method for organometallic mercury removal is the spheric propane. Propane refrigeration at the interstage of the
IFP/Procatalyse process which, through hydrogenolysis of low temperature mixture refrigerant compression system has
the organometallic compounds at moderate conditions on a also been suggested for helping to balance the compressor
catalyst bed reactor, yields metallic mercury. The elemental 60 power between the propane and mixed refrigerant systems.
mercury thus produced is then trapped at a lower temperature Other proposals include the use of vacuum propane and
in a second reactor on a bed of the Procatalyse CMG-273 ammonia absorption refrigeration.
catalyst. The number of Heat Transfer Units (HTU) in a heat
New Mercury Treatments: A number of new treatments exchanger is defined as the temperature drop of the warm
show promise for helping meet mercury waste specifications. 65 stream (or temperature rise of the cold stream) divided by the
Physical separation technologies based on the high density of average temperature difference between the warm and cold
mercury are used to treat contaminated soils and have some streams. Kettle type heat exchangers have a limited number of
US 8,381544 B2
11 12
HTU while plate fin, spiral wound and other exchangers stream 446 from the debutanizer 440 can be mixed with the
whose warm and cold flows are counter-current have greater condensate from the inlet/upstream section or can be used as
limits on HTU. a gasoline component.
The use of a multicomponent refrigerant for precooling Traditionally the removal of NGL from the feed gas has
excludes the use of conventional kettle type heat exchangers. been integrated into the precooling section, which provides
Aluminum plate-fin heat exchangers have been proposed for partial condensation for separation, an embodiment of this
this duty. Other types of heat exchangers have also been arrangement is shown in FIG. 1. However there are cases
Suggested for use in LNG plants such as Heatric, Packinox, where front end NGL recovery, which is upstream of the gas
and UOPHigh Flux Tubing. Another practice involves the use 10
cooling section, may be the preferred design. A study was
of hydraulic turbines instead of expansion valves on the LNG conducted by KBR to compare the efficiency of front end
product and liquid mixed refrigerant streams. The advantage versus integrated NGL recovery for an LNG production facil
is due to the isentropic expansion of the liquid, with energy ity in a neutral approach to liquefaction technology. The study
recovery, compared with isenthalpic expansion across a will use a generic liquefaction process as a basis for compari
valve. 15 son. The analysis is not based on any proprietary liquefaction
Liquid Removal: Heavy hydrocarbon components (e.g., process. The NGL recovery technique varies with the place
aromatics and Cs and higher aliphatic hydrocarbons) can be ment in the process. Front end NGL recovery will use a
removed via line 212 to prevent them from freezing in the conventional expander plant design with full pressure recov
liquefaction section 230. Lighter components such as ethane ery. Integrated NGL recovery can employ different schemes,
and liquefied petroleum gas (LPG, comprising primarily pro Such as condensation by refrigerant or expander technology.
pane) can be removed for one or more of the following pur A series of process simulations were used to model the liq
poses: to provide make-up refrigerant, to control the heating uefaction process with selected NGL recovery methods at
value of the LNG product, to prevent freezing of aromatics in various temperature levels in the feed gas chilling train. A
the liquefaction process, or to allow separate sale of gas comparison based on specific power was applied to evaluate
liquids as a product. The feed gas stream can be cooled Such 25 the options. The results of this study can serve as a useful
that the heavy hydrocarbon components (e.g., NGL) are con guideline for future process design. The configuration of the
densed and then can be removed by vapor/liquid separation in NGL recovery system by necessity changes somewhat as its
a trayed column. The trayed column is sometimes referred to placement within the chilling train varies. Two concepts were
as a scrub column. If this column is re-boiled, the scrub 30
used as the basic approaches to NGL recovery for all of the
column can also function as a demethanizer. Further separat cases considered. The first employed a high pressure column
ing of the liquid stream can be carried out in the fractionation with external refrigeration to generate reflux for the required
section 220. The temperature to which the gas is cooled separation. The second used a traditional "expander plant”
depends on the amount of LPG that is desired to be recovered. design utilizing feed gas expansion and recompression to
Fractionation: The liquids recovered from the natural gas, 35 provide the necessary refrigeration for the recovery cycle.
often comprising ethane through pentane, are called natural This approach was further varied by looking at the impact of
gas liquids or NGLs. The hydrocarbons in NGL can be recov eliminating recompression.
ered from the Scrub column and separated in a fractionation In all cases, the refrigeration power was kept constant
train that comprises, for example, demethanizer, deethanizer, while the production rate was allowed to vary. This approach
depropanizer, and debutanizer columns. The bottoms stream 40 is consistently used in many facilities where the refrigeration
from the demethanizer can be fed to the deethanizer. The compressors are driven by a fixed selection of gas turbine
demethanizer can be eliminated if the upstream scrub column drivers. The power consumed by the recompressor, although
is re-boiled. The demethanizer overhead vapor can be included in the process specific power, is not considered as
returned to the natural gas feed stream or can be sent to the part of the refrigeration compression. The significance of this
fuel gas system for the plant. FIG. 3 depicts a schematic 45 distinction is that it allows additional power to be applied to
illustration of a NGL/LPG fractionation unit. In FIG. 3, one the process that can have an impact on the overall production
embodiment of a fractionation train 400 to separate an inlet rate. The end use of the recovered NGL can vary from site to
stream 412 is shown comprising a deethanizer 402, depro site. In some cases, the recovered NGL will be exported as a
panizer 420, and debutanizer 440 columns. The bottoms separate product; in others, the recovered NGL is re-injected
stream 406 from the deethanizer can be fed to the depro 50 into the LNG product after satisfying refrigerant makeup
panizer 420, and the bottoms from the depropanizer 426 can requirements. In this study, the recovered NGL is not returned
be fed to the debutanizer 440. The deethanizer overhead 404 to the LNG product stream; therefore, the specific power,
can be returned to the natural gas stream, up to a desired limit. which is compression power referenced against LNG product
Alternatively, ethane can be collected for sale as an external rate, will inherently rise with increasing NGL recovery. The
product or can be used as fuel in the plant. A side draw stream 55 results of this study are that the optimum design of an NGL
408 from the deethanizer column top section can be taken as recovery system within an LNG production facility depends
liquid ethane to the refrigerant make-up system. This stream on the target objectives.
does not necessarily require stringent distillation because the Low Specific Power/High Efficiency: From an efficiency
make-up refrigerant can be a mixture with methane and pro standpoint there is a clear trend Supporting an integrated NGL
pane. The bottom stream 406 from the deethanizer should be 60 recovery system with a relatively cold feed. The data indi
very lean on ethane if the downstream depropanizer is used to cated a significant improvement in overall system perfor
produce refrigerant grade pure propane (e.g., 99.5% pro mance with a -40°C. feed to NGL recovery compared to
pane). warmer draw points. There was virtually no change in specific
Propane 428 and butane 448 can be recovered as liquids power from a front end design (40°C. feed) to a -15°C. draw.
from the total condensers of the depropanizer 420 and debu 65 Lower levels of ethane recovery offer the highest efficiency,
tanizer 440 columns. These liquids can be used for refrigerant which is as expected. Higher levels of NGL recovery con
make-up or sent to LPG storage via line 430. The bottoms sume additional power and reduce the LNG product stream.
US 8,381544 B2
13 14
When the specific power is measured relative to total liquid tions and without causing freezing problems, then the feed
product, the overall trend is much less sensitive to the ethane gas can pass through the LNG plant at the Supercritical pipe
recovery level. line pressure. This can improve the efficiency of the liquefac
High Production Rate: Under the basis of design condi tion process and can simplify the facilities by eliminating
tions laid out in this study, high efficiency is not necessarily NGL recovery and its liquid handling facilities. However, this
synonymous with high production rates. While refrigeration alternative would not allow for recovering refrigerant make
power is restricted to a fixed value, recompression power up fluid without having an expander-compressor inclusion
associated with NGL recovery is unconstrained. This pre and could require importation of refrigerant.
sents the opportunity for additional power to be applied to the With higher liquefaction pressure, less power is required
overall system, which can have a positive impact on produc 10 by the liquefaction process. The higher pressure can also
tion rates. This is best illustrated by noting that an NGL allow a higher outlet temperature from the main heat
recovery feed temperature of -40°C., which corresponded to exchanger, which can result in a richer or heavier mixed
the best efficiency point, yielded the lowest production rate at refrigerant (MR) that can reduce refrigeration power require
all levels of ethane recovery. Varying the feed temperature to ments or increase the throughput.
NGL recovery from 40° C. to -15° C. had relatively little 15 The refrigeration circuit can extract energy from the natu
impact. LNG production was highest at low ethane recovery ral gas and reject the energy to the environment. In some
levels, while total liquid production was highest at high embodiments, one, two, or three refrigerant circuits can be
ethane recovery levels. This demonstrated that although used, involving pure refrigerant, mixed refrigerant and/or a
increasing NGL recovery pulled potential product away from combination of pure and mixed refrigerant. Base load lique
the LNG stream, increasing NGL recovery allowed for pro faction processes generally use mechanical refrigeration, in
cessing a higher feed flow, which resulted in more total liquid which heat is transferred from the natural gas through
production. For those processes where recompression power exchanger Surfaces to a separate closed loop refrigerant fluid.
was not used, the highest product rate naturally corresponded The circulating refrigerant provides the necessary cooling.
to the highest efficiency point since the total power applied to The refrigerant fluid is colder than the natural gas through
the system remained constant. 25 compression and pressure let-down expansion. The types of
Optimum System: As indicated in the sections above, this refrigerant systems that can be used include pure refrigerant
study did not indicate an optimum approach that was equally and mixed refrigerant (MR). One difference between pure
valid for all facilities. The constraints placed upon each refrigerant systems and mixed refrigerant systems is the step
design can shift the design philosophy in different directions. wise or cascading cooling effect of pure refrigerant versus the
There does appear to be a best efficiency point that would 30 evaporation curve of a mixed refrigerant paralleling the con
support a colder feed to the NGL recovery system. There also densation curve of the natural gas as shown. FIG. 4a depicts
appears to be an opportunity at warmer feed temperatures to a graph showing the condensation curve of a mixed refriger
apply additional power through recompression that can ant system. FIG.4b depicts agraph showing the condensation
increase product yields in refrigeration constrained systems. curve of a nine level cascade pure refrigerant system. The
Liquefaction: The natural gas can be cooled by mechanical 35 natural gas condensation can take place in one piece of equip
refrigeration. A refrigerant gas can be compressed, cooled, ment with intermittent withdrawal and reentry, or in two or
condensed, and let down in pressure through a valve that more exchangers in series, for example. Plate fin and spiral
reduces its temperature by the Joule-Thompson effect. The wound exchangers can allow several streams to be cooled in
refrigerant gas can then be used to cool the feed gas. Some the same piece of equipment. In one embodiment, the refrig
constituents of the natural gas, such as propane, ethane, and 40 eration is accomplished by a propane pre-cooled MR process
methane can be used as refrigerants, either individually or in that combines the cascading effect of the pure refrigerant with
a mixture. Suitable types of liquefaction cycles include cas a mixed refrigerant as a second cooling media. One embodi
cade, mixed refrigerant, and expansion cycles, as well as ment of a method having a multi-component refrigerant com
combinations of two or more thereof. The compressors used bined with a cascade method is U.S. Pat. No. 4,404,008 to
in the refrigeration system may be driven by Steam turbines, 45 Rentler et al., which is incorporated by reference herein.
gas turbines, electric motors or combinations thereof. Designing for normal steady state conditions is important
The dry lean gas is condensed in a cryogenic exchanger, but unusual operating conditions should also be considered. A
sometimes referred to as a cold box or MCHE (main cryo few important examples are: varying cooling water (or air)
genic heat exchanger) to an outlet temperature that will result temperature to the condenser, changing refrigeration load;
in complete liquefaction of the feed gas. In at least some 50 and compressor recycle. Referring to FIGS.5 and 6, these two
situations, the liquefaction pressure can be kept as high as systems may seem equivalent in operation and the capital cost
possible and the outlet temperature can be kept as warm as and normal steady state operating costs of these two systems
possible to improve liquefaction efficiency. The exchanger are essentially the same. The difference between the two is in
outlet temperature can be close to -161° C. without subse how they will respond to the examples of unusual operating
quentendflash or liquid expansion. Plate fin (PF) and spiral or 55 conditions listed above.
spool round (SW) exchangers can be used and can allow a FIG. 5 depicts a schematic illustration of a conventional
close temperature approach. They are also very compact com refrigeration system for an LNG facility. FIG. 6 depicts a
pared to the shell & tube exchangers. schematic illustration of an alternate refrigeration system for
The choice of operating pressure for the liquefaction plant an LNG facility. Both the conventional system 600 shown in
can affect other aspects of the process and apparatus. Higher 60 FIG. 5 and the alternate system 602 shown in FIG. 6 contain
operating pressure reduces the refrigeration load, but the the same major components of a compressor 610, compressor
pressure can remain significantly below the cricondenbar (the driver 612, desuperheater/condenser 620, accumulator 630,
highest point of the phase envelope, or the highest pressure at process heat exchangers 640, 642, suction drums 650, 652
which two phases are still possible) to allow vapor/liquid and one or more lines 660 to enable the flow of refrigerant to
separation. 65 other stages of refrigerant use.
If the composition of the feed gas is such that the NGL can Varying Cooling Medium Temperature: The cooling
be retained in the LNG while still meeting product specifica medium temperature to the refrigerant condenser 620 can
US 8,381544 B2
15 16
have large variation depending on the season, weather con exchanger 650/640, there is a significant liquid piping inven
ditions, and with air cooling the temperature changes from tory on the downstream side of the valve. When the recycle
day to night can have a significant effect. In the conventional begins, the liquid piping inventory is forced into the drum/
scheme (FIG. 5), colder cooling medium than normal results exchanger 650/640 creating a sudden undesirable level rise.
in the lowering of the compressor 610 discharge pressure due The obvious Solution is to design the system to minimize the
to the excess surface area in the condenser 620; to control the liquid inventory between the control valve 632 and the drum/
suction pressure above vacuum conditions, the driver 612 exchanger 650/640.
speed is reduced. This control scheme works quite well in Another solution to the problem of compressor recycle and
most situations. However, if the cooling medium temperature liquid inventory control is the use of vapor recycle (without
variations are large, or if a driver 612 with a fixed or limited 10 liquid cooling of the vapor). The heat of compression is
range of speed is employed, the scheme 602 shown in FIG. 6 removed via a desuperheater on the compressor discharge.
is more suitable. In this design, a control valve 622 is installed This solution avoids concerns about liquid inventory and
between the condenser 620 and the refrigerant accumulator liquid carryover to the compressor because refrigerant liquid
630. The control valve 622 can flood liquid into the condenser cooling is not employed. The primary disadvantage of this
620 (reduce surface area) to increase the compressor 610 15 scheme, however, is that during the recycle operation, the
discharge pressure which in turn increases the Suction pres compressor sideload temperature rises above the normal
sure; this is especially useful with a fixed speed driver 612 operating temperature; but this system is acceptable since the
because it avoids the potential for vacuum Suction conditions compressor can be designed for the full range of operation.
due to a low discharge pressure. Compressors and Drivers
Changing Refrigeration Load: Normally the level of refrig Reducing the number of machines per train while increas
erant in the accumulator 630 and the compressor suction ing the train capacity is the Subject of considerable investiga
drums 650, 652 remain relatively steady, however, when the tion. A recent development, associated with high capacity
process load changes, the level in the accumulator 630 should LNG trains, involves the use of large gas turbines to drive
easily respond to the varying demand for liquid required by multiple refrigerant compressors connected in tandem on a
the evaporative exchangers without causing secondary dis 25 single shaft. The large gas turbines can require a starter, which
turbances in the control scheme. The system 602 in FIG. 6 can be utilized for other duties when not used as a starter and
accomplishes this objective. The pressure of the accumulator in some instances have been used as power generation units.
630 can be controlled at a lower pressure than the compressor The starter can be used to supplement the power of the gas
610 discharge pressure via the control valve 632 that can vent turbine during normal operation if desired. Other possible
vapor from the accumulator 630 to the high level refrigerant 30 configurations include: a combined cycle where waste heat
sideload pressure (note that the control valve 622 between the recovery from gas turbine(s) generates steam utilized by a
condenser and the accumulator, as discussed in the previous steam turbine driving another compressor; large electric
section, is also necessary in order to “decouple' the compres motor drivers associated with inexpensive hydroelectric
sor 610 discharging pressures from the accumulator 630 oper power or as part of a specific powerplan development project;
ating pressures). Thus, with the ability to vary the vapor 35 and/or the use of available pressure in the feed gas stream to
generation in the accumulator 630, the level in the accumu drive an expander and isentropically lower the temperature of
lator 630 is easily varied without disturbances to the com the feed using the power from the expander to supplement the
pressor 610 operation. With the scheme of FIG. 6, the accu refrigeration systems.
mulator 630 can then more easily accomplish its main Liquefaction Processes
function intermediate storage of liquid inventory that 40 The following is a representative listing of liquefaction
responds readily to refrigeration demand. processes that have been developed: APCI Propane pre
Compressor Recycle: When the compressor 610 is in cooled mixed refrigerant, C3MR, DUAL MR; Phillips Opti
recycle operation, the compression energy must be removed mized Cascade: Prico single mixed refrigerant; TEAL dual
by the desuperheater/condenser 620 to avoid temperature pressure mixed refrigerant; Linde/Statoil multifluid cascade:
build-up in the refrigeration system. In many refrigeration 45 Axens dual mixed refrigerant, DMR; and Shell processes
systems, such as 602 shown in FIG. 6, the liquid formed in the C3MR and DMR.
condenser 620 is directly injected into the vapor recycle Heat Transfer Equipment
(desuperheating) at each stage to maintain the proper tem Shell and Tube Units: The main use of shell and tube type
perature at that refrigeration stage. This method works very exchangers in LNG plants is for cooling water service and for
well under many operating conditions. The primary short 50 boiling refrigerant services down to temperature levels of
coming, however, is that due to either poor temperature con -50 F. Below a temperature of -50 F., alloy construction
trol or the limited desuperheater operating range, too much such as nickel alloys, for example 3.5% and 9% nickel, stain
liquid may be injected at one or all stages and result in exces less steel, or aluminum is required. It is seldom that the shell
sive liquid accumulation at the lower level compressor Suc and tube exchanger remains competitive when temperatures
tion drum 652 or exchangers 640, 642. This problem can be 55 are colder than -50 F. when a large number of HTU are
resolved by the direct injection/sparging of the recycle vapor required. The significant savings in exchanger Surfaces per
via control valve 632 into either the drum 650 or evaporative mit a reduction in hardware, making these units economical.
exchanger 640 as shown in FIG. 6. This system is operable Aluminum core (plate-fin) exchangers: The plate-fin
throughout the entire recycle range possible with the com exchanger is typically made up of heat exchange Surfaces
pressor 610 and is more reliable and predictable. Note that in 60 obtained by stacking alternate layers of corrugated, die
general, the preferred method of sparging is into the formed aluminum sheets (termed fins) between flataluminum
exchanger 640 if possible because sparging into the compres “separator” plates which can vary in thickness from 0.032 to
sor suction drum 650 increases the drum size and creates 0.064 inches, depending on design pressure. The separator
undesirable frothing. Sparging into the heat exchanger 640 sheets Supply the primary Surface and the fins Supply the
more easily handles these concerns. An interesting design 65 secondary, extended Surface. In normal design practice, the
detail is the exact placement of the vapor injection control secondary surface varies from 67% to 88% of the total heat
valve 632. If the valve 632 is located too far from the drum/ transfer Surface provided in an exchanger. Each layer is
US 8,381544 B2
17 18
closed at the edges with solid aluminum bars of appropriate individual streams and also have to pass the pressure test,
shape and size. A full size heat exchange is made by stacking fabricated core sizes must be reduced with increasing design
many layers, one on top of the other. This multi-layer Stack is pressures. Generally, maximum sizes at design pressure of
then bonded together by a carefully controlled vacuum braz approximately 650 psig are 3 ft.x3 ft.x16ft. and the heat
ing process to yield an integral rigid structure with a series of 5 exchange Surfaces per core are generally limited to approxi
fluid flow passages. mately 40,000 sq. ft.
The dimensions of the corrugated aluminum sheets (fins), For liquefaction plants, considerably more heat exchange
which actually form the fluid passages and provide the Surface is required than can be Supplied by a single core and,
extended heat exchange surfaces, can be varied widely with therefore, it becomes necessary to combine a large number of
respect to quantity, shape, spacing size and type, dependent 10 cores requiring extensive manifolding. A typical unit process
upon both thermal and hydraulic design as well as manufac ing 300 million std. cu. ft./day may require 30 to 40 cores,
turing economy. Fins are normally furnished with Straight dependent on process conditions. Ensuring somewhat equal
corrugations in “plain”, “perforated' or “lanced’ configura flow through each core can be a significant challenge. These
tions. Fin height can he varied from 0.200 to 0.355 in., metal manifolded core assemblies are generally located within a
thickness from 0.008 to 0.025 in., and fin density from 6 to 25 15 cold box together with other equipment and piping as part of
fins/in. The actual selection of the most suitable fins for any the low temperature process section. In order to avoid exten
particular application is therefore, dependent upon the maxi sive field assembly, involving aluminum welding, complex
mum working pressure, plus other variables such as heat fit-ups, etc., core assemblies may be shipped to a seaside
exchange rates, allowable pressure drops, fluid properties, assembly point where the final construction and assembly of
and fluid flow rates. Each of the several different fluids being the cold boxes can be performed. These assembled cold boxes
handled simultaneously in a given exchanger is accordingly (excluding insulation) may then be shipped as a unit by barges
assigned a certain passage geometry also based on a careful or other suitable water transportation means. The cost for
optimization of calculated thermal and hydraulic perfor manifolding, coldbox fabrication, and assembly is significant
mance. Then the different passage geometries are stacked up and may more than double the cost of the actual cores. Nev
similar to a sandwich, alternating symmetrically into one of 25 ertheless, the overall installed cost of such a unit is competi
the several flow patterns available. The most common of these tive with other types of cryogenic exchangers.
used in process plants is the counter flow pattern gas-to-gas Spiral Wound Exchangers: The spiral wound heat
and liquid-to-gas. exchanger as the name implies, is made up of tubes which are
Boiling or condensing applications often use a cross-flow wound on a mandrel, as thread or cable is wound on a spool.
pattern that minimizes pressure drop in the exchanger, an 30 The exchanger will have a tube sheet at either end to which
important process consideration in the boiling stream. Fol each tube is joined. Normally, a layer of tubes is wound (say
lowing the brazing and cleaning of the exchanger core, the left to right) on a mandrel, and spacers (bars, wire, etc.)
collectors and nozzles can then be welded onto the unit. Then, attached to the tube. This is followed by a second layer wound
a pneumatic test of at least 110% of the design pressure and/or in the opposite direction (right to left), and then a third (left to
hydrostatic pressure test of at least 130% of the design pres 35 right again), each layer complete with its own set of spacers.
Sure is typically applied. Brazed aluminum heat exchangers This procedure is repeated until the required number of tubes
usually comply with the ASME Pressure Vessel Code and are has been wound onto the mandrel. For the large exchangers
usually so certified. Plate-fin exchangers are most useful in used in LNG plants, tubing diameter will usually range from
complex process cycles because it is possible to accomplish % to 3/4 inch in diameter. These tubes, which can range from
within a single exchanger unit the same heat exchange that 40 75 to 150 feet in length, can be applied to the mandrel with a
would ordinarily require multiple two-passage shell and tube winding angle or helix of approximately 10°. This means that
exchangers. a layer of /2 in. tubes at 5 ft. diameter with a pitch of 0.75 in.
Plate-fin exchangers can easily be designed to process five will consist of 40 parallel tubes, while a layer at 10 ft. diam
or six streams and it is possible to withdraw and add streams eter would consist of 80 parallel tubes.
as required along the length of the heat exchanger. This fea 45 The tube winding is generally been performed on gigantic
ture makes it possible to accomplish a partial condensation of lathes. As bundle sizes increased, additional temporary Sup
a feed stream within the heat exchanger, then remove the feed ports can be used to minimize the deflection of the mandrel
stream, separate the liquid and re-introduce the vapor back due to the bundle weight causing detrimental movements
into the exchanger for additional condensation. There is great between tubes and spacers during the winding operation. The
flexibility offered to designers by using plate-fin exchangers. 50 winding can also be performed in the vertical position,
Plate-fin exchangers are normally considered for operation in thereby minimizing the weight effect of the bundle. The tubes
processes that are non-corrosive to aluminum, non-fouling, can be made of aluminum or copper or possibly other mate
and free of particles that could plug fin spacing of 15 or 16 rials. When using aluminum tubing, aluminum tube sheets
finsfin. Most cryogenic processes meet these criteria. and shells are generally used, although, in some cases, it may
Because most warm feed streams are coming from a pipeline, 55 be more economical to use stainless tube sheets and shells
it is considered a wise precaution to use filters or strainers with aluminum tubes.
before entering a plate-fin exchanger. Even traces of com By using more than one tube sheet at either end it is pos
pressor oil may build up on the fin Surface over long periods sible for these exchangers to handle more than one fluid in the
of time and foul or plug the exchanger. tubes. By using a very long mandrel, it is possible to wind
The size of individual cores is dependent on fabrication 60 several exchangers on a single mandrel and enclose them all
facilities and operating conditions, primarily pressure. The in a single shell. This feature is particularly useful when the
size of the brazing furnace and the ability to evenly distribute process employs the same shell side fluid for several exchang
heat during the brazing process sets maximum limitation with ers in series because this feature greatly reduces the need for
respect to the size of cores. The exchangers can be, for field construction labor. Generally, the LNG spiral wound
example 4 ft.x4 ft.x20 ft. long with heat exchange Surfaces 65 exchanger units are mounted vertically with the cold section
approaching or exceeding 100,000 sq. ft. per core. Since at the top. The high pressure streams are inside the tubes. The
individual cores have to withstand the full design pressure of vaporizing low pressure refrigerant flows downward in the
US 8,381544 B2
19 20
shell side. Although the technology for fabricating spiral LNG Product Pumps: The LNG product pump has a special
wound exchangers is well known, the advent of large LNG design for cryogenic service. The pump is a Submerged
plants has dramatically increased the available sizes for this motor, “pot mounted pump for these applications. The con
type of exchanger. Individual preferences, process or site tainer, flooded with LNG during operation, also contains the
conditions, or economics may dictate the use of one or the motor. The suction of the pump is at the bottom of the con
other of the exchanger types mentioned above. tainer, and the LNG discharge flows through the motor thus
Since most exchanger applications within the cryogenic providing cooling for the motor. This arrangement does
section involve boiling and condensing, the specific hydraulic require cryogenic rotating seals; the only seal needed is for
problems of two-phase flow require special attention with the electrical connection box, and the box is always purged
respect to flow direction, distribution, and mixing. These 10 with nitrogen to prevent natural gas leakage through the con
requirements may favor Some exchanger types or may require duit. LNG Product Pumps has at least the following advan
special design details. It is not practical to reach general tages over conventional sealed pumps: (i) the LNG Product
conclusions with respect to optimum selections. Heat Pump is completely Submerged in the pumped fluid, resulting
exchanger requirements for each sending terminal can be in reduced noise; (ii) the LNG Product Pump does not contain
extensive and regardless of which type exchanger is being 15 rotating shaft seals that are difficult to design and maintain for
selected, a major strain can be put on the available worldwide the cryogenic temperatures; therefore, inflammable gas is not
fabrication facilities which may have a significant effect on leaked into the atmosphere (the pump does have static Seals in
the delivery schedule. This is particularly true if several large the electrical conduits to seal around the main power Supply
LNG plants are being considered at the same time. Since and instrumentation wiring); (iii) the LNG Product Pump
these exchangers are not solely used for LNG terminals, but uses a single shaft design with both the pump impellers and
are also specified for any other low temperature process, their motor on the same shaft, eliminating the need for a coupling
availability is dependent on the overall demand of cryogenic and removing alignment issues; (iv) the LNG Product Pump
plants. New concepts in either design or fabrication, with motor and pump bearings are product lubricated, eliminating
particular emphasis on extremely large units, most likely the need for an external lube oil system; and (v) the LNG
being fabricated of aluminum are therefore desired. 25 Product Pump does not require an explosion proof motor.
Liquefaction Pump Services As an operational and design challenges for a Submerged
Examples of major pump services in the liquefaction unit motor LNG pump, the Suction pot of the pump must be liquid
includes: Amine circulation (acid gas removal process); filled prior to starting the pump. Cool down of the pump is a
Reflux for scrub column and fractionation towers (liquefac delicate activity that must be done slowly to prevent excessive
tion process); LNG product pumps; Seawater pumps (if sea 30 thermal stresses and damage within the pump. Various meth
water cooled); and Hot Oil Pumps. ods are used to try and ensure that the pump is properly cooled
Amine Pumps: The amine pumping service is often split down and liquid filled prior to start-up. These include moni
into two parts: a low head pump working at high temperature toring the temperature on a vent/bleed connection to the
followed by a high head pump operating at near-ambient pump, use of temperature sensors within the pump Suction
temperature. Using the low head booster pump at the high 35 container and a level gauge on the Suction pot. Further, in a
temperature avoids problems with cavitation within the pump cryogenic application, condition monitoring is difficult since
which would be present if the high head pumping were done the vibration monitors need to be placed inside the cryogenic
at high temperature. The booster pump is typically a single Suction pot mounted on the pump. Some other options that
stage double section pump with low NPSH (Net Positive have been used are external vibration instruments on the
Suction Head) requirements. By using a pump with low 40 cover plate of the Suction pot and operating without vibration
NPSH requirements for the booster pump, the residual dis instrumentation. The LNG Product Pumps have historically
solved CO, remains in solution. When CO, is allowed to been very reliable, therefore for many users, operation with
come out of Solution, a phenomenon similar to cavitation out condition monitoring instrumentation has been an accept
occurs that is potentially very damaging to the pumps. To able solution.
avoid the potential for cavitation damage, calculated NPSH 45 Seawater Cooling vs. Freshwater Loop Cooling: The sea
available numbers are typically reduced by three to four times water pumps are very large in a base load LNG plant and the
to provide Sufficient actual margin. The amine circulation rate pumps are typically mounted vertically in a seawater intake
depends on the amount of acid gas, but a train making 5 basin. The flow rates of these pumps are commonly in the
MMTPA of LNG with a natural gas feed containing 15% CO, 15,000 to 18,000m/hrrange. The head may vary between 50
can have a circulation rate over 2000 cubic meters/hour 50 to 60 meters. Large, Vertical, open pit, multi-stage pumps are
handled with 3x50% pumps. The high-head circulation commonly used. In some plants the seawater removes heat
pumps are typically multi-stage, between bearing, horizontal from a fresh water loop, instead of the more common once
designs driven by electric motors. through cooling where the seawater goes directly through
Reflux Pumps: The reflux pumps for the scrub column heat exchangers and then discharges back to the sea. The fresh
operate at about -30°C. to -50° C., and in the fractionation 55 water loop circulation rate is similar to the seawater rate, but
unit the deethanizer reflux pumps also operate at about -30° the liquefaction unit exchangers exchange heat with fresh
C. The flow rates of these pumps depend to a large extent on water. The advantage of using the extra cooling loop is higher
the natural gas composition. For a 5 MMTPA train handling reliability and lower cost materials in the liquefaction unit.
associated gas the scrub column reflux flow can be in the 350 The disadvantages are extra cost and equipment for the fresh
to 400 cubic meters/hour range, though a plant processing 60 water loop and a higher heat sink temperature for the process
non-associated gas usually has a smaller Scrub column reflux (which makes the process require more energy). Fresh water
pump. The scrub column reflux pump size depends to a great circulation pumps are normally horizontal, double Suction
extent on the aromatics present, but in some cases where the designs.
natural gas contains little ethane and propane, recovering Hot Oil Pumps: The liquefaction process, in spite of being
refrigerant components can be the main factor that deter 65 cryogenic, still requires some heating services. Examples are
mines reflux pump size. These pumps are normally single the amine stripper reboiler and fractionation reboilers. How
Stage. ever, most gas turbine driven LNG plants do not have heat
US 8,381544 B2
21 22
recovery steam generation (HRSG), and in Such cases hot oil Fuel Gas System
is a common heat transfer medium. The hot oil is circulated The fuel gas system receives Supply from the feed gas, flash
between the heat source and process services with hot oil gas at the back end of the liquefaction train, storage and
pumps. In some cases, steam is used as aheating medium, and loading area, and miscellaneous sources Such as the fraction
in Such cases condensate pumps and boiler feed water pumps ation area, etc. To minimize the overall power requirements
replace the hot oil pump services. This substitution com for liquefaction, the flash gas system provides the majority of
monly takes place when there are enough sulfur compounds the fuel gas requirements for the plant. If the flash gas com
in the gas to make Sulfur recovery in a Claus unit necessary; pressor trips off, fuel make-up is from the feed. This system
the Claus unit generates low pressure steam which is available operates well as long as the heating value of the feed and the
for process heating services. Another option that has been 10 flash gas are close. However, in plant designs containing a
Successfully used is to incorporate a waste heat recovery unit significant amount of nitrogen in the feed, the flash gas also
in the exhaust of the gas turbine and utilize a heated water contains significant nitrogen; for example, 5% nitrogen in the
circuit for heating. The heated water system is maintained feed gas results in about 40% nitrogen in a typical flash gas
under pressure to prevent boiling, and a centrifugal pump is system. The primary concernis that the flash gas and the feed
used for circulation. Hot oil and hot water circulation pumps 15 gas are not immediately interchangeable at the feed users
can vary widely in design, but horizontal double Suction (heaters, boilers, etc.).
designs are commonly used. Hot oil pumps typically have a A system can be designed to "ramp' the change in heating
capacity between 1500-2000 cubic meters/hour, and a head value from predominantly flash gas to all feed gas so that the
between 120-140 meters. Heated water pumps typically have fuel burner controls at the users can be adjusted at a reason
a capacity between 750-1250 cubic meters/hour, and a head able rate; without an appropriate system, the heating value
of 220-250 meters; water has a higher heat capacity than oil, change to the user could be essentially a step change (see FIG.
hence circulation rates tend to be smaller. 8). FIG. 8 is a graph illustrating the changing heating value of
Cryogenic Liquid Expanders: One other service in lique fuel gas in a step change and a ramping change. One method
faction related to pumps is the cryogenic liquid expander as to accomplish this ramping of heating value uses a fuel gas
an alternative to a JT (Joule Thompson) valve. The liquid 25 mixing system that is capable of controlling the rate of change
expander (or hydraulic turbine) is like a pump running in to a manageable level (for example, 20% to 30% change per
reverse; the fluid enters at high pressure and exits at lower minute) as shown in the desired ramp in fuel heating value in
pressure, and shaft power is generated instead of being con FIG.8. FIG.9 depicts a schematic illustration of one embodi
Sumed. The drop in pressure is controlled with a back-pres ment of a mixing vessel to control heating value changes in
Sure valve to prevent the discharge from flashing into two 30 fuel gas. FIG.9 depicts one embodiment of a fuel gas mixing
phases. Two different technologies have been used for the system 700 providing a linear ramping of the heating value
cryogenic liquid expander application. The first is a sub and includes a mixing vessel 702 with a fuel inlet 704, and
merged motor LNG pump operating as a liquid turbine. For fuel outlet 706, internals such as a center tube 708 for an exit
this design, the expander/generator speed is controlled by stream connected to the fuel outlet 706 and baffling 710 to
using a Variable Speed Drive System (VSDS). This design 35 create multiple chambers within the vessel 702. Many lique
has the advantages of the mechanical portions of the LNG faction plants do not need Such a mixing system since the
cryogenic pumps, i.e. no seals and couplings. The second heating value differences between the flash gas and the feed
approach is to use a liquid expander similar to a vertical gas do not require ramping. Other solutions may also be
turbine pump. This technology requires the use of a shaft seal possible. Such as for example, heating value measurement
(either dry gas or oil film) and an external generator. The 40 combined with feed forward control on the boiler alone or in
performance of the turbine is controlled using a set of wicket combination with a fuel gas mixing system may be feasible on
gates (inlet guide vanes) to control the pressure drop across Some projects, depending upon the circumstances.
the expander. The speed of the expander is fixed with the LNG Storage
synchronous generator connected to the electrical grid. An important item in LNG facilities is LNG storage. In
Nitrogen Rejection (Endflash): FIG. 7 depicts a schematic 45 Some embodiments, in order to minimize cost, it can be useful
illustration of one embodiment of an endflash unit for nitro to maximize the size of each LNG storage tank. Described
gen rejection. In FIG. 7, illustrating one possible embodi below are a few types of storage tanks. Use of higher pressure
ment, the endflash section 500 can remove nitrogen from the storage tanks can eliminate use of blowers for vapor control.
LNG. After liquefaction of the natural gas at high pressure in Careful layout design can also reduce piping costs. The use of
the liquefaction section 510, the LNG pressure can be 50 large, below ground tanks may offer a more economic solu
reduced, such as through one or more valves 512, 514 to tion where plot space is limited. The storage tanks are
approximately atmospheric pressure before entering the Stor equipped with relief valves as a defense against overpressure.
age tanks 526. This eliminates high vapor generation in the Vacuum breakers can provide protection against external
tank that would have to be recompressed by a boil off gas overpressure.
compressor, utilized as fuel, flared, or otherwise utilized. An 55 Single Containment Systems
endflash 500 can be used if the nitrogen content in the LNG is FIG. 10 depicts two illustrated examples of single contain
above about 1%. The endflash 500 also can remove methane ment LNG storage tanks. In FIG. 10, the inner wall or primary
with the nitrogen that can be returned to the fuel gas system by container 60 of the single containment tank can be con
re-pressurizing the gas to a fuel gas pressure. The endflash structed of a material, such as a 9% nickel steel, which can
section 500 can comprise a flash drum 516 and/or a re-boiled, 60 contain the refrigerated liquid and can be self-supporting.
trayed column 520 for more extensive nitrogen removal. The This inner tank can be surrounded by an outer wall 62 which
column 520 can concentrate the nitrogen and reduce the can be of a different material, such as carbon Steel, that can
methane loss from the LNG. The vapor can be routed through hold insulation, such as perlite, in the annular space between
an exchanger 522 to recover most of the cold energy before the inner and outer walls 64. A carbon steel outer tank 62 is
being compressed in the fuel gas compressor 524. Column 65 not capable of containing LNG, thus the only containment is
520 can also be a flash drum instead of a trayed column, in that provided by the inner tank 60. The base can have insula
which case the exchanger 518 may be eliminated. tion 66 and some embodiments can have a suspended deck
US 8,381544 B2
23 24
roof 68 that can also be insulated. Single containment tanks capability to introduce LNG into the top or the bottom section
are surrounded by a dike 70 or containment basin external to of the storage tank. This allows mixing LNG of different
the tank, either of which provide secondary containment 72 in densities and can reduce rapid vapor generation. Filling into
the event of failure or leakage of the LNG. Embodiments can the bottom section can be accomplished using an internal
have external insulation 74 and can have bottom heating 76 to standpipe with slots, and top filling can be carried out using
prevent freezing the ground and causing heaving. In some separate piping to a splash plate in the top of the tank.
embodiments the tanks can be elevated above grade. Such as Seismic Requirements: Seismic design requirements are
utilizing an elevated concrete raft structure, which can pro dependent on the geographic location for the tank. Some
vide additional room for spill containment and eliminate the locations have good quality data available for specifying the
need for bottom heating. 10
seismic design requirements. Other locations may have a
Double Containment Systems history of seismic activity but the quality of data available
FIG. 11 depicts two illustrated examples of double con may not be comprehensive. This may lead to over or under
tainment LNG storage tanks. In FIG. 11, double Containment specification for the earthquake intensities for the operational
systems include a secondary wall 78 that is capable of con
taining both liquid and vapor. The inner wall 60 can be con 15 basis earthquake (OBE) and safe shutdown earthquake (SSE)
structed of a material, such as 9% nickel steel, which can cases. At a seismically active location a large diameter tank
contain the refrigerated liquid and can be self-supporting. The with a low height to diameter ratio may result in a more
roof 68 over the inner tank can be carbon steel. Double con economical design compared to a design utilizing seismic
tainment tanks have an outer wall 78, such as a steel or isolators. Installation of seismic isolators requires two base
concrete wall, capable of holding LNG. In Double Contain slabs and may reduce the horizontal and vertical seismic
ment systems no dike is needed because the outer wall pro design forces by 50%. For example, the tanks at Marmara
vides the secondary containment for the LNG. LNG vapors, Eregisi terminal in Turkey have a large diameter with low
however, may be released in the event of an inner tank leak in height design, whereas those at the Revithoussa terminal in
systems where there is no sealed roofto the outer wall. A roof Greece have seismic isolators under the tank base.
80 that is not sealed to the outer wall 78 can be provided and 25 Tank Design Pressure: Full Containment tanks have been
an earth embankment 82 can be placed exterior to the outer designed for pressures up to 300 mbarg, whereas Single and
wall 78. Double containment tanks have been traditionally designed
Full Containment Systems for 150 mbarg maximum and usually much less. Higher oper
FIG. 12 depicts two illustrated examples of full contain ating pressures of Full Containment tanks with concrete roofs
ment LNG storage tanks. In FIG. 12, a Full Containment 30 allows the installation of lower capacity vapor handling
system includes a secondary wall 78 that is capable of con equipment for ship unloading, which can result in Substantial
taining both liquid and vapor that has roof 80 over the outer cost savings. All metal tanks have been built with 250 mbarg
wall. Such as a concrete or steel roof, making the outer tank design pressure. The additional cost from increasing the pres
capable of handling both LNG liquid and vapor. The inner sure from 150 mbarg to 250 mbarg was reported to be
wall 60 can be constructed of a material, such as a 9% nickel 35 approximately 2% of the tank cost. The higher design pres
steel, which can contain the refrigerated liquid and be self Sure capability gives a distinct advantage in reducing the cost
supporting. The roof 68 over the inner tank 60 can be carbon of vapor handling equipment. However, with Single Contain
steel. If the inner tank leaks, all liquids and vapors can still be ment or other bare steel roof higher pressure designs, the high
contained within the outer wall 78 and roof 80. There can be cost of long pipe runs, deluge water and foam equipment can
insulation 84 on the inside of the secondary wall 78. 40 not be eliminated.
Membrane Systems Condensation in the Annulus of a Double-Walled Cryo
FIG. 13 depicts two illustrated examples of membrane genic Storage Tank
LNG storage tanks. In FIG. 13, a Membrane system utilizes a Inflow of Vapor into Annulus: The absolute pressure in the
membrane material capable of containing the LNG. The annulus is essentially the same as in the vapor space of the
membrane type storage tank can be a pre-stressed concrete 45 tank. Therefore, any condensed vapor should be replaced by
tank with a layer of internal insulation covered by a mem inflow of vapor into the annulus. If this incoming vapor has
brane. Such as a thin stainless Steel membrane, that is capable the same composition (and hence the same dew point) as the
of containing the LNG and serves as the primary container 60. condensing vapor, condensation may continue. But if the
In this case the concrete tank 78 supports the hydrostatic load incoming vapor has a dew point temperature lower than the
which is transferred through the membrane 60 and insulation 50 inner-tank-wall temperature the vapor will not condense in
(in other words, the membrane is not self-supporting or load the annular space. Considera case where the annulus contains
bearing). The membrane can shrink and/or expand with pure methane. If LNG containing nitrogen is then introduced
changing temperatures. into the tank the methane vapor in the annulus will condense
FIG. 14 depicts two illustrated examples of cryogenic con because of the colder LNG temperature. However, the boil
crete LNG storage tanks. The primary container 60 can be 55 off vapor from the LNG will be a mixture of nitrogen and
constructed of cryogenic concrete that is designed to with methane, with a dew point temperature lower than that of
stand the cold temperatures of LNG service. The secondary methane. This nitrogen-methane mixture will replace the
wall 78 can be constructed of pre-stressed concrete and can condensed methane vapor in the annulus, and further conden
have a carbon steel liner 86. sation will stop. For this case the total amount of condensa
FIG. 15 depicts two illustrated examples of spherical LNG 60 tion will be limited to the amount of pure methane in the
storage tanks. The primary container 60 can be enclosed Volume of the annulus. The condensed methane will re-va
within an outer shell 88 that in some embodiments can be porize due to heat leak. Because of buoyancy forces the
partially buried or covered with an earthen berm 90. methane vapor will flow up the face of the outer wall and
The common industry practice is to have all connections to eventually escape from the annulus. For a period of time a
the tank (e.g., filling, emptying, venting, etc.) through the roof 65 methane-rich Zone will exist near the bottom of the annulus
so that in the event a failure of a line should occur the failure and this may create a cycle of vaporization and recondensa
will not result in emptying the tank. Each tank can have the tion.
US 8,381544 B2
25 26
Warming of The Tank Contents: When condensation able gas (or gas mixture) in the annulus can ensure that a low
occurs in the annulus, there is heat transfer from the annulus enough dew point is always maintained.
to the liquid in the inner tank. A temperature driving force First Stage LNG Loading Pumps: Low-head pumps can be
across the inner-tank wall is therefore essential. As the tank located in each LNG storage tank. These pumps can operate
contents warmup, the available temperature driving force and fully submerged in LNG, and can be located within pump
the rate of condensation will decrease. To illustrate this case wells or columns for easy installation and removal without
consider an LNG tank filled with pure liquid methane satu taking the tank out of service. The pump wells can also serve
rated at one atmosphere. The tank boil-off vent is then closed as the discharge piping for the pumps and can be connected to
until the tank pressure increases by 50 mbar. When this hap the tanktop piping. These pumps can deliver the desired LNG
pens, the methane dew point temperature rises by 0.6°C. If 10 send-out flow and can also circulate LNG through the ship
the pressure rise occurs within a short period the temperature loading piping to keep the lines cold between times when
driving force of 0.6°C. will be available immediately, and ships are being loaded. In an embodiment, a suitable dis
charge pressure for an in-tank pump can range from about 50
condensation will begin. The condensed methane will be psig to about 200 psig. In an alternate embodiment, a Suitable
replaced by more methane from the tank vapor space, and 15 discharge pressure for an in-tank pump can range from about
condensation will continue So long as there is an adequate 100 psig to about 150 psig. In an alternate embodiment, a
temperature driving force across the wall. As the condensa Suitable discharge pressure for an in-tank pump can range
tion proceeds, the liquid in the tank will gradually warm up from about 100 psig to about 125 psig.
because the condensation process will give up heat to the Two types of loading pumps are a vertical pump with
inner tank. In addition there will also be normal heat leak Submerged motors, and vertical-shaft, deep-well pumps with
through the tank wall, floor, and roof. For 100,000 m of pure externally mounted motors. Both types can be used and alter
liquid methane a Sub cooling of 0.6° C. represents enough natively or additionally multistage horizontal pumps can be
refrigeration to condense over 165,000 kg (390 m3) of liquid used. Vertical pumps with Submerged motors are most often
in the annulus. For a 50 m diameter tank with a 1 m wide chosen.
annulus, this condensed methane would represent a pool of 25 Vertical Pump: A vertical pump with submerged motor can
liquid about 2.5 m high. For a 50,000 m pure ethane or be constructed in Such a manner that the pump with motor
propane storage tank of 40 m diameter the condensation drive is hermetically sealed in a vessel and Submerged in the
corresponding to 0.6°C. Sub cooling would be 1.2 m and 1.3 liquid being pumped. The major advantage of this design is
m of liquid in the annulus, respectively. These figures, of that the extended shaft with its associated seal is eliminated.
course, have been derived from an oversimplified model, and 30 Since the problems with most cryogenic pumps lies in the
it is unlikely that operators would permit Such a deep pool of dynamic seals, eliminating them may provide a more reliable
condensate to be formed. For a pure component stored in a design. This type of design has the pump and motor surround
double-walled cryogenic tank, an increase in pressure can ings 100% rich in LNG, and thus would not support combus
result in severe condensation in the annulus. When the liquid tion. Also any ingress of moisture is stopped and problems
in the tank is a mixture, the dew point of the vapor in the 35 due to differential shrinkage of materials is reduced or elimi
annulus will depend upon the vapor composition, and the nated. In this design the LNG itself cools the motor windings
severity of condensation will be less than for pure compo and lubricates the motor bearings. This type of pump may be
nentS. used in ship loading and unloading applications and for
Decrease in Tank Pressure: A decrease in tank pressure will pumping of LNG out of LNG storage tanks. In some embodi
cause a drop in the dew point temperature of the vapor in the 40 ments utilizing a high head Submersible pump can eliminate
annulus. The net result is a decrease in the temperature dif the need for second stage LNG send-out pumps.
ference across the tank wall which in turn causes the conden Vertical-Shaft Pump: A vertical-shaft pump is configured
sation to either decrease or to stop. This technique is instan with an externally mounted motor connected to a pump by a
taneously effective for pure components. shaft, requiring a seal between the pump and shaft. The seal
Methods to Prevent Annular Condensation: Methods to 45 can be a mechanical seal. A vertical-shaft deep-well pump
prevent condensation can be selected by establishing the cri with an externally mounted motor can be used for LNG
teria that are necessary for condensation to occur, and then service, but can pose safety concerns regarding the possibility
operating the facility so that condensation is avoided. Con of failure of the mechanical seal on the extended shaft and
densation in the annulus can be prevented by ensuring that at possible exposure to LNG vapors to the externally mounted
any given tank pressure the vapor in the annular space has a 50 motor. If the first stage send-out pumps are located inside the
dew point lower than the temperature of the tank contents. tanks, they will likely be of the submersible design. If they are
Except in the special case of a stratified tank, the temperature outside the tanks, however, then they will most likely be a
in the tank will nowhere be expected to be colder than the considerable distance from the tanks; that is, the unloading
bubble point temperature at the tank operating pressure. The pumps will be located out of the confines of the diked area,
dew point temperature of the annulus vapor can be controlled 55 and the risk of exposure to LNG vapors is reduced, thereby
either by controlling the composition of the vapor or by making the use of a vertical-shaft pump feasible.
changing the pressure in the tank (and hence in the annulus), LNG Storage Pumps: In one embodiment the pumps are
but in practice absolute pressure control may not be a suitable inside pump columns located within the storage tank that
method. Usually there will be other criteria that establish the extend to the storage tank roof The key design feature of this
operating pressure in the tank So that pressure is not normally 60 pumping system is that it is possible to pull the pump for
available as a parameter for dew point control. However, as maintenance while continuing to operate the storage tank.
pointed out later, under Some situations pressure reduction There can be a foot valve at the bottom of the column that
may be the only available method for lowering the dew point. prevents LNG from entering the column when the pump is
Effective control of the dew point temperature is best pulled. The operators purge the column with nitrogen, and
achieved by composition control in the annulus. The annulus 65 then remove the pump from the top of the column. The LNG
is normally dead-ended and filled with perlite and fiberglass loading pump capacities are often based on filling a ship in
insulation. Hence a small continuous purge of a non-condens twelve hours. The liquefaction plant typically can have mul
US 8,381544 B2
27 28
tiple storage tanks and 2 to 4 pumps per tank. It is common to vacuum or near vacuum conditions, and then sealed. The heat
have a total of eight pumps running during loading, each with leakage from this system can be substantially less that of the
a capacity in the 1100-2000 cubic meters/hrrange and 150 typical mechanical types of insulation. Under special circum
240 meters of head. In many plants there is also a smaller stances it may be worthwhile to design a piping system that
pump in each tank in addition to the loading pumps. The has two structural barriers capable of containing the LNG.
purpose of this smaller pump is to recirculate LNG in the This may be accomplished in several ways, such as for
loading lines and stabilize the temperature when no ship is example, the vacuum-jacket piping may be designed Such
present. The loading lines can be large diameter (for example, that the outer pipe is also suitable for cryogenic temperatures.
24 inch to 36 inch) and are typically kept cold between ship Alternatively, the piping may be installed within a cold box
loadings because cooling them down is a long procedure. 10 that is constructed to withstand the internal and external
The pumps used for the in-tank application can be similar forces. For example, a concrete cold box could be installed;
to the LNG product pumps except they are mounted in a the cold box could be filled with bulk insulation, sealed and
column connected to the top of the tank instead of in a vessel. pressurized.
The pumps can use Submerged motors that are cooled by Safety
passing the LNG product flow past the windings of the motor. 15 A real-time plant management system known as the LNG
Special care must be taken when the pumps are removed from Plant Advisory System (LNGPASTM) has been developed at
the tank because the winding insulation can be very hydro KBR to guide and assist the LNG Plant Operators in the safe
scopic and can absorb moisture. Nitrogen purging of the control and operation of the LNG storage facility. LNGPAS is
pumps is recommended when they are not in use. The condi an advanced process control system which combines conven
tion of the pumps can be monitored by using accelerometers tional programming techniques with knowledge based sys
mounted on the pump housing close to the bearings. The tem technology to reduce the complexity of monitoring and
pump bearings are typically a stainless steel material and controlling the operation of an LNG storage facility. The
lubricated by the LNG product. Reliability of the foot valve is system receives on-line process data from existing sensors,
as critical as the reliability of the pump. The foot valve is analyzes them, detects abnormalities, and advises LNG plant
required to seal when the pump is removed to allow the tank 25 operators on any corrective actions. For example, the events
to remain in service and is typically Supplied by the pump which may lead to a potential tank content “rollover situa
Supplier as an integral part of the pump. The weight of the tion or the effects of any corrective actions recommended to
pump sitting on the foot valve causes the foot valve to open prevent a “rollover situation can be predicted and analyzed
and allows LNG to enter the pump and column pipe. using the built-in process simulator.
Insulation 30 Some type of alarm prioritization system or other operator
Insulation: Some of the basic types of insulation used for advisory system has become a necessity for safe plant opera
LNG plant piping are mechanical types or vacuum jacketing. tion. The lack of such a system may result in an operator's
Within the mechanical types there is also the distinctions of alarm saturation syndrome which could develop into a poten
pre-insulated VS. field-insulated; and polyurethane VS. cellu tially hazardous situation owing to lack of operational
lar glass such as FOAMGLAS(R) from Pittsburgh Corning 35 response to incubation of rollover leading to dangerous
Corporation. Many LNG facilities use polyurethane due to its events. This is particularly true in LNG terminals where
good thermal conductivity and because polyurethane is rela operators are continuously involved in the movement of large
tively economical. However, since polyurethane is less stocks of product and concurrently handling the vapors gen
impervious to vapors than FOAMGLASR), provisions must erated by those transfer operations. In these LNG terminals, a
be made to ensure that a good vapor barrier is provided to 40 safe design demands a large number of alarms that require the
protect the insulation from deterioration due to water ingress. operators attention; some immediately and others in the
It is also important to design the insulation system such that shorter or longer term. With time and the operators increased
combustible gas does not leak from the piping into the insu familiarity with the plant, however, many of these alarms tend
lation because this may present a hazard. FOAMGLAS(R) is to lose their sense of urgency, especially those which are not
advantageous in that FOAMGLAS(R) is impervious to water 45 very frequent or do not require immediate attention. Under
vapor, thus it is easier to protect against insulation deteriora these conditions, potentially dangerous situations may arise,
tion due to water ingress. FOAMGLAS(R) also has a higher which could be avoided if timely preventive measures were
compressive strength than polyurethane, which can resultina implemented. One typical example of such a situation is when
more durable application. gas of a given composition and light density is added to a
Preinsulated piping offers advantages because it mini 50 storage tank partially filled with gas of a different composi
mizes field labor and because production-line manufacturing tion and heavier density. The formation of two layers of LNG
can in Some instances increase quality control. The major product inside the tank may result owing to inadequate mix
disadvantage of preinsulated pipe, aside from cost, is the ing. The mixing of the Stratified layers is accomplished by a
possibility of shipping and Schedule delays. Preinsulated pipe massive/rapid increase in vaporization rate that could develop
is usually shipped to the facility site with the ends left bare. 55 into a hazardous situation. This hazard can be minimized by
The pipe can then be welded and the ends are then field adopting safe operating procedures for loading LNG into
insulated via preformed rigid insulation or the insulation can storage tanks, as well as by taking the necessary corrective
be field applied in the manner referred to as poured-in-place. actions to assess the case of a potential rollover in an incu
In general it is preferred to use preformed rigid insulation for bating tank. Real-time knowledge based systems are ideally
larger piping because there can be problems associated with 60 suited for these kinds of problems. The process knowledge
large pours. entered into the system as rules can validate and analyze
Vacuum-jacket piping may also be considered for LNG sensors/alarms, identify trends, diagnose abnormalities, rec
facilities. This type is constructed such that there are two ommend or take corrective actions, and so on.
piping walls; the inner wall that is constructed of a material to Rollover: Filling a cryogenic LNG storage tank with pro
contain the LNG and an outer wall that may be constructed of 65 duction or shipments of different densities may result in the
carbon steel or other material. The annulus between the two formation of stratified liquid layers (fill-induced stratifica
piping walls can be filled with insulation, evacuated to form a tion). Once stratified liquid layers are formed, the heat-leak
US 8,381544 B2
29 30
into the tank from the Surroundings into the bottom layers is monitors the filling and sendout operations of all storage
not released and is stored as Superheat. This is owing to the tanks. Operator decides on whether to use top or bottom
inability of the natural convection currents in the bottom layer filling procedures when filling a tank and which tank to use
to penetrate entirely through the top layer and come to the free for filling, sendout or any other transfer operation based on
Surface to release heat as vapor. As a result, the density of the density, composition, and temperature data. A process simu
bottom layer decreases with time owing to its increasing lation system can be used to simulate dynamically the events
temperature. At the same time, the top layer is continuously leading to potential rollover or to analyze the effects of any
able to release its heat in-leak by vaporization of lighter corrective actions recommended to prevent rollover. The pro
components, which results in a continuous increase of its cess simulation system predicts incubation time and performs
density owing to concentration of heavy components. A roll 10 overall heat and material balances on each tank and the whole
over can occur when the density difference between the two storage area. The alarm management system monitors alarms
layers becomes sufficiently small that the natural convection and trips in the storage area. The alarm management system
currents from the bottom layer come to the free surface. The determines the events which may lead to an alarm or trip
Subsequent mixing of these layers is accompanied by a large activation, established their urgency level, and recommends
increase in the normal vaporization rate, which will be pro 15 type of operator intervention. The maintenance and diagnos
portional to the amount of Superheat accumulated in the bot tics system monitors key operating and mechanical param
tom layer. This physical phenomenon associated with the eters for machine performance, analyzes them and correlates
mixing of stratified layers of LNG is commonly and descrip data to provide preventive maintenance.
tively referred to as “rollover. If several layers of stratifica Stratification detection: Several methods are available to
tion develop at a given time, the same phenomena may occur detect stratification in LNG storage tanks, such as density
among the different layers. In that case, roll over of bottom measurementalong the height of the tank; temperature profile
layers may develop without heat release and may continue along the height of the tank; measurement of changes in the
until only two layers remain. Similarly, the top two layers vaporization or boil-offrates; and/or measurement of changes
may rollover with the corresponding vapor release, Subse in the boil-off composition.
quent to which the tank remains stratified with the new top 25 Typically, the most reliable measurement is the density
layer and the remaining existing bottom layers. along the height of the liquid in the tank. This is normally
Stratification and Subsequent rollover are not always unde provided by a probe that can travel the full height of the tank
sirable. Generally, the contents of a well-mixed tank roll over and routinely measures the density at different levels. The
continuously with low vapor release. A rollover of high inten number and frequency of these measurements can be changed
sity can result, however, in vapor release in excess of the 30 as desired in order to determine as closely as possible the
designed vapor handling capabilities of the tank and can thus existence of stratification and the thickness of the stratified
overpressure and possibly rupture the tank. The resulting layers. These measurements are recorded routinely at the
hazardous scenario must be avoided and all precautionary shortest practical time intervals. If stratification is detected,
steps should be taken to prevent stratification as well as to the reading/measurement intervals can be repeated as fre
defuse the stratification once stratification occurs. This roll 35 quently as necessary, at the same time, temperatures along the
over phenomenon is not exclusive to LNG: rollover can also height of the tank can also be measured and monitored regu
occur in liquefied petroleum gas (LPG) stored at atmospheric larly to detect Sudden changes in temperature profile which
pressure, and even in pure materials such as ammonia, ethyl may indicate stratification. Temperature profiles are not as
ene, or ethane, usually with very minor release of vapors if reliable as density profiles in detection of stratification but are
initial temperature Stratification develops as a result of for 40 helpful when combined with density profiles. The real-time
instance, addition of a warmer product on top of a colder one. knowledge-based system primarily monitors the density and
The phenomenon of self-stratification is also possibly a result temperature profiles of the tank to detect stratification. The
of the weathering process of an initially homogeneous tank. system also calculates the heat and material balance of the
For instance, high concentration of nitrogen in the LNG storage/loading area and predicts the normal boil-off rate
(about 4% or higher) can result in self-stratification and mod 45 which is then compared with the actual boil-off as a further
erate rollover, because the boil-off from the free surface check in case of large deviations for the detection of stratifi
decreases the nitrogen concentration in the upper layer of the cation.
stored liquid, which makes the upper layer lighter than the Stratification Detection System: This system monitors the
bottom liquid. Rollover can generally be prevented by fol density and temperature profiles along the height of the Stor
lowing good operating procedures. It is possible that, owing 50 age tank. The system compares the density and temperature
to a number of causes (operator error, lack of available storage measurements at a given height with the density and tempera
or proper top and bottom filling devices, etc.), stratification ture measurements at the adjacent level immediately below
may still develop in a given tank. When stratification is and makes a determination whether the tank is stratified at
detected, corrective action can be taken, Such as recirculation that level or not. These comparisons are performed at every
from bottom to top or bottom to bottom of the tank after 55 level for which the density and temperature measurements are
releasing its Superheat by flashing vapors or accelerated send available, starting from the bottom of the storage tank and
out—to avoid the hazardous consequences of a serious roll continuing until the maximum liquid level is reached. The
OVer. system performs these comparisons at fixed time intervals,
Rollover risk can be minimized by using the following when new sets of density and temperature profiles are made
sub-systems: Stratification Detection System; Filling/Send 60 available. The system also maintains historical trends of these
out System; Process Simulation System; Alarm Management profiles for future reference and the system notifies the plant
System; Maintenance and Diagnostics System, The stratifi operator whether or not stratification was detected at every
cation detection system monitors the density and temperature level for which the density and temperature comparisons are
profiles along the liquid height of each storage tank. The made. If stratification is detected, the system can estimate the
stratification detection system compares these measurements 65 thickness of stratified layers and calculate their average den
and notifies the operator when stratification is detected and sity and temperature. These values are then used to estimate
advises on any corrective action. The filling/sendout system the rollover incubation time and intensity, using appropriate
US 8,381544 B2
31 32
parametric equations, and/or by performing a series of pro high gas Velocity and resultant high heat transfer rate at the
cess flash calculations. At this point, the system advises the nozzle; and/or depressuring a distillation column with light
plant operator on any corrective actions and recommends the liquid hydrocarbons present may result in freezing the reboil
start-up of one or more re-circulation pumps to minimize the ing media (condensate, hot water, etc.). In this case, proper
rollover intensity and extend the incubation time. The system feed forward instrumentation for quick draining of the heat
also recommends a recirculation mode (bottom to top or ing media or a method to maintain continuous flow of the
bottom to bottom) based on the expected intensity of rollover. media may be necessary.
LNG Storage Tank Filling/Unloading System: This system FlareSizing: Since the depressuring system is a significant
monitors the filling, unloading or standby operation of a load on the flare system, the interaction of the two systems is
storage tank to minimize or prevent stratification. The system 10 important in the overall integrated design. Flare sizing should
advises the plant operator on whether to use top or bottom fill consider the impact of the depressuring system upon the
procedures when filling a storage tank. This determination is dynamics and thermal shock of the flare system and should
made based on the density difference between the LNG in the avoid simultaneous release of depressuring circuits under
storage tank and that of the incoming LNG to ensure adequate operator control. Flare sizing should optimize the size of the
mixing. The system estimates the tank boil-off or vaporintake 15 depressuring circuits and their resulting load on the flare
rate, by taking into account filling or unloading flow rates, system; for example, the multi-staged propane refrigeration
boil-off owing to heat leak from the Surroundings, pumping system may be divided into separate circuits via Solenoids and
energy, vapor generated due to liquid flash, etc. The system trip valves to ensure the depressuring of two individual cir
also provides a variety of other useful information, Such as cuits on a staggered basis. Flare sizing should ensure that the
total volume filled or unloaded, time required to fill or unload depressuring circuit does not relieve faster than design (usu
a tank and so on. ally fifteen minutes). This may require the use of travel stops
Process Simulation System: This system consists of a built on depressuring valves.
in process simulator and database with the ability to perform Emergency Shutdown (ESD) and Emergency Depressur
different types of flash and heat boundary layer penetration ization Systems (EDS):
calculations. The system can be used to simulate the events 25 The safety of personnel, plant equipment and environment
that may lead to potential rollover situations and to determine is achieved in part by the implementation of an emergency
how the density and temperature of the stratified layers may isolation system and an emergency depressurization system
change with respect to time based on the composition of which is activated in case of fire, potentially dangerous pro
various layers, recirculation rates, and heat leak from the cess upsets, or hydrocarbon leakages. The process plant area
Surroundings. The system can also be used to see how the 30 can be divided into possible fire Zones with sectional plot
rollover incubation time and intensity may be affected based areas containing equipment with a given maximum hydrocar
on operator corrective actions. The simulation results can be bon inventory. Proper arrangement of the process equipment
presented in graphical or tabular format. LNGPAS was devel should be considered during the plot plan design that could
oped using the Realtime Advisory Control (RTAC) System result in comparable Volumes of hydrocarbon C or lighter
environment by Mitech Corporation. The system also uses 35 liquids per each fire Zone. Each Zone can be isolated at its
KBR’s specialized process simulator to perform heat and boundaries by the emergency shutdown valves before pro
material balance calculations. The current system can be ceeding to depressurization of equipment. Depressurization
adapted to various equipment configurations and enhanced to is the rapid reduction of process equipment pressure by
handle multiple tanks or provide full heat and material bal relieving its inventory to flare or vent. This is particularly
ance calculation capability. The system can be used as an 40 important for a vessel exposed to fire. Relief valves are
operator training tool and also as an alarm prioritization sys designed to keep vessels below their design pressure, but not
tem capable of determining the event(s) leading to alarms, to reduce the pressure. As fire increases the metal temperature
establishing their urgency level and recommending timely thus reducing the material strength, lowering the vessel pres
operator intervention. The LNGPAS is a user-friendly menu sure reduces the stress on the metal, which reduces the risk of
driven system that allows the operator to switch from one 45 the vessel bursting, therefore reducing or preventing further
system to another and access sensor data or other pertinent damage to the plant. Depressuring rates are proposed in API
information. 521.
Depressuring System: In addition to the conventional pres There are other potentially dangerous situations where it is
Sure relief system, a depressuring system is generally pro desirable to remove the process fluid inventory from the pro
vided to reduce the internal pressure of equipment either 50 cess equipment to a save destination, for instance, equipment
involved in or adjacent to a fire. An overall depressuring close to an area on fire. In Such cases the plant may be
system can be segregated into many independent circuits that depressured through either vapor or liquid depressuring
can be discharged into the flare system on a controlled basis. valves to suitable flare or vent facilities. The depressuring
Thermal effects and flare sizing are important design consid process results in a rapid isentropic expansion of the vessel
erations. 55 content as the depressuring performs work on the relieved
Thermal Effects: Remotely operated vapor depressuring fluid. This will cause drastic reduction on the temperature of
valves are provided as per API RP521 for equipment/piping the fluid in the vessel particularly when depressuring mix
systems containing more than two tons of liquid hydrocar tures of low boiling point hydrocarbons. Since heat transfer
bons (butane or lighter). The depressuring of hydrocarbons, between the vessel and its contents can reduce the vessel
however, results in significant temperature reductions that 60 metal temperature to below the ductile/brittle transition tem
must be carefully analyzed. For example, the lower tempera perature, depressuring can severely reduce the stress on the
ture may determine the materials of construction; the forma equipment. Therefore, the depressuring scenario often deter
tion of liquid from isentropic vapor depressuring and the mines the minimum design temperature of the process equip
continued depressuring of the liquid may result in significant ment.
temperature reduction in the equipment; the gas entering or 65 The depressurization philosophy adopted in the design is a
leaving a vessel during depressuring may significantly lower critical factor in pertaining to metallurgy selection in an LNG
the nozzle design temperature, relative to the shell, due to the liquefaction plant. The main criteria are whether the repres
US 8,381544 B2
33 34
surization is allowed while cold or not. If an immediate tors such as personnel safety, process safety, accessibility,
repressurization is possible, impact test qualified low tem operability and maintainability have to be considered care
perature and cryogenic grade materials will be required in fully in a facility layout as these issues are typically inter
most portions of the liquefaction and fractionation trains. If related. Often the final facility layout is a compromise among
controlled repressurization is adopted, not allowing repres these various but inter-related factors.
Surization while the equipment or piping is at cold liquid One way to achieve a safe and economic layout is by
temperature, the use of fine carbon or low alloy steels will studying different layouts using three-dimensional Software
predominate in the plant design. In the former case the cost of and estimating the thermal and vapor dispersion profile on
the plant increases significantly but impact test qualified low each one of the different layouts. This procedure can be done
temperature and cryogenic grade materials provide additional 10 quite fast if variations of the layout are carried out by simple
safety in case an operator does, in fact, repressure the plant re-plotting of the main equipment which is then followed by
after a plant upset that resulted in depressurization. In the computer re-running of the pipe-racks and production of the
latter case the cost of the plant is reduced, but the operator is cost estimate for the new piping arrangement.
not allowed to restart for an immediate repressurization while Instrumentation
the plant equipment or piping is at cold liquid temperature, 15 The measurement of operating conditions, such as for
thus reducing plant availability. It is important, however, to example, pressure, temperature, flowrate and liquid level are
have adequate temperature measurements to ensure safe important in LNG plants. For cryogenic applications, how
repressurization. ever, special instrument design details may be necessary to
Liquefaction Plant Layout ensure that the measurements are accurate and reliable. The
Overall layout of the plant, i.e. location of the storage area, accurate measurement of liquid level in a vessel via a differ
the process area, utility area, loading area, flares, control ential pressure (D.P.) cell can be critical in an LNG liquefac
room, etc. have to take into consideration not only the oper tion plant.
ating units inside the battery limits but also the communities Level Measurement via Differential Pressure (D.P.) Cell:
outside the plant boundaries. Separations between above FIG. 16 depicts a schematic illustration of a conventional tap
areas should allow for effective fire fighting and to avoid fire 25 into a vessel. An upper tap 800 penetrates the vessel wall 802
from one area to propagate to others. Each area should have and vessel insulation 804, the upper tap outside of the insu
access from at least two different ways. Equipment with a lation 804 can have an upward slope 806 and proceed to a
high inventory of flammable material which could develop differential pressure transmitter. A lower tap 810 penetrates
large vapor clouds in case of leakage should be located down the vessel wall 802 and vessel insulation 804, the lower tap
wind of the prevalent wind direction away from community 30 outside of the insulation 804 can have an upward slope 806
areas, control rooms, warehouses, etc. Electric power to the and proceed to a differential pressure transmitter. The differ
plant should be provided through at least two separate feeder ential pressure transmitter can measure the difference in pres
circuits and the fire water system should be looped around the sure between the upper 800 and lower 810 taps. FIG. 17
entire plant such that any fire water unit can be supplied from depicts a schematic illustration of an alternate thermosyphon
either direction. 35 tap arrangement. FIG. 17 demonstrates an arrangement that is
In developing an overall layout, calculations of the thermal more suitable for measuring the level of LNG. A thermosy
radiant profile and vapor dispersion contours produced by phon arrangement on the liquid level taps can be used to
code specified spill rates have to be checked to ensure com ensure constant vapor formation at point A. Instead of a single
pliance with the applicable codes. As an example, NFPA 59A lower tap there are two, a first lower tap 812 and a second
code specifies that provisions shall be made to minimize the 40 lower tap 814. The first lower tap 812 can have a downward
possibility of the damaging effects of fire or a flammable slope 816 to junction point A, whereas the second lower tap
cloud of vapors from Such a design spill to reach beyond a 814 can have an upward slope 818 to junction point A, where
property line that can be built upon and that would result in a the first lower tap 812 and the second lower tap 814 connect.
distinct hazard. Therefore, the battery limits of the LNG After connecting at point A the lower tap can have an upward
facility may be set by the above calculated vapor and thermal 45 slope 806 and proceed to a differential pressure transmitter as
radiant contours. Today, the possibility of a hostile attack on in the conventional system. The connection and slopes of the
an LNG facility is of concern and a generic assessment of the first lower tap 812 and the second lower tap 814 create a
worse case consequences resulting from a deliberate action thermosyphon arrangement on the lower taps to ensure con
against the facilities should be considered. Of course, the risk stant vapor formation at point A. The circulation through the
of such an event must be weighed in terms of the severity of 50 thermosyphon is facilitated by the ambient heat leak that is
the consequences as well as the probability of occurrence. controlled by the proper combination of insulation thickness,
The assessments of the consequences have to be backed by an tube diameter and slope. This technique may solve the two
evaluation of historical, experimental and theoretical evi main problems that exist with the conventional approach: a
dence. The results of the analysis can then be included in the stagnant tube encourages the plating out of Solids (such as
hazard footprint to identify those areas at risk from gas cloud 55 carbon dioxide, benzene, etc.) present in the LNG that can
dispersion or radiated heat from fire. Similarly, ground level result in tube plugging; and a rising level in the tube creates
concentration of gases released from vent and/or flare have to vapor and a resultant higher pressure that provides a false
be calculated in order to ensure proper concentration levels pressure signal from the bottom level tap.
beyond the plant boundary limits. These problems may be solved by the constant circulation
At the same time, process area layout of equipment should 60 and continuous vapor generation that is achieved in the sys
be arranged to minimize piping between equipment, to pro tem shown in FIG. 17. There are other techniques such as the
vide area for the cleaning and maintenance of equipment, and use of a bubbler (introduction of an external bubbling gas)
to provide depressurization in case of fire or to protect nearby that may also resolve the above-mentioned problems but may
equipment from an existing fire. Different areas within a unit not be as reliable or as simple as the design provided above.
should be properly curbed and drained to avoid the possibility 65 Shell and Tube Exchangers
of pooling of flammable material under equipment and to Materials of Construction: Normally the selection of mate
direct spills to designated impounding areas. Numerous fac rials is straightforward. The major area of concern for
US 8,381544 B2
35 36
exchange services near-46 degrees Celsius is the selection of out dirt, Scale, etc. that may clog up the relatively small
the proper minimum design temperature. Since the design passages within a plate fin exchanger.
break point between impact tested carbon steel and stainless Orientation/Configuration: Due to the exchanger size limi
steel is -46 degrees Celsius, the impact of depressurization tation, parallel units are typically required; thus, flow balanc
(including tube rupture) and the resultant temperature drop 5 ing between units is important. Proper distribution of a two
must be carefully analyzed to make the correct materials phase stream into a plate fin exchanger is essential. Generally,
choice. a horizontal configuration is not recommended because of the
Thermal Stresses: As a rule, to minimize thermal stresses potential separation of vapor and liquid within the exchanger
on a fixed tubesheet exchanger, limit the temperature differ due to the effects of gravity; thus, a vertical configuration is
ence between the shell and tube to about 20-40 degrees Cel 10 preferred. The “coldend up' vertical configuration is possible
sius. The actual allowable stress on the exchanger is evaluated and generally results in reduced pressure drop (power Sav
for each exchanger. Just as for the materials of construction ings) due to taking advantage of the liquid hydrostatic head in
selection, a design temperature analysis is necessary to deter downflow. During shutdown, however, an undesirable tem
mine temperature variations during normal operation as well perature inversion occurs in the exchanger due to the cold
as during start-up, shutdown, depressurization, etc. For 15 liquid at the top of the exchanger settling to the bottoms. For
example, in normal operation, the propane level on the shell this reason, this operation is not recommended for plate fin
side of a feed/propane kettle exchanger must always be above exchangers. Generally, the “cold end down vertical configu
the tubes. If not, the tubes above the propane level are warmed ration is preferred. When the cold liquid settles to the
to the feed temperature which results in tube expansion and exchangerbottoms on shutdown, there are no severe tempera
possible excess stress at the tubesheet and the shell/tubesheet ture inversions that cause excess thermal stress. Also, the
interface. exchanger can be restarted quickly due to the stable tempera
Leaks: A leak analysis is recommended before the ture profile.
exchanger design is finalized to assure that the exchanger type Two-Phase Flow Distribution: If the inlet vapor volume
and details are suitable for the service intended. A leak analy percentage is less than 10% or more than 90% then special
sis should include examination of the potential cause of leak, 25 inlet distribution is not needed. Between 10% and 90% vapor
the probable direction of leak, the effect of leak on process Volume, external separation of vapor and liquid and the forced
operating conditions, the method of detection in operation, distribution of each phase into the exchanger via special inter
and the recommendations on leak prevention, repair, nals is recommended.
exchanger type, and necessary design details. Innovative Heat Exchanger Designs: Enhancements of the
Orientation/Configuration: With a few exceptions, conven 30 conventional shell and tube heat exchanger Such as the butted
tional orientation and configurations are employed, but in tube sheet design have been effectively used in existing LNG
Some cases modular engineering techniques are applied to plants, and, along with other modifications such as enhanced
save on heat exchanger and piping costs. An example can be surfaces and better materials of construction, will find
a modular exchanger design for feed and mixed refrigerant expanded use in the future. In addition, several new heat
cooling via propane refrigeration. The process fluid can flow 35 exchanger designs have been developed to address the limi
without interruption from one exchanger to the next, thereby tations in commonly used heat exchangers. Examples of these
the normal exchanger heads and connecting piping can be new exchanger designs include, but are not restricted to:
eliminated on several of the exchangers. Heatric (an exchanger made from etched plates fused
Two-Phase Flow Distribution: The primary area where together to maintain full metal strength throughout the
two-phase flow distribution is important for shell and tube 40 exchanger, available from Heatric, a Meggitt group company
exchangers is in the feed and mixed refrigerant cooling via of Dorset, UK with a sales office in Houston, Tex.): Packinox
propane forecooling. With the conventional arrangement, (an exchanger made by welding sheets that have been formed
there is separation of the tubeside liquid and vapor between by an explosion technique, wherein the plate forming tech
exchangers with the liquid flowing preferentially through the nique reduces residual stress in the material making the mate
lower bank of tubes. The net result of this separation is to 45 rial less prone to corrosion attack, available from Packinox
increase both the overall heat transfer surface and the refrig S.A. in France, an Alfa Laval company); and High-Flux and
eration power. The modular butted tubesheet arrangement Fine Fin (can provide enhanced Surfaces to promote high
discussed above overcomes both these drawbacks. performance via higher heat transfer rates).
Plate Fin Exchangers Suitable experience at cryogenic temperatures will be
Materials of Construction: These exchangers are made of 50 needed before these new exchanger designs will be consid
aluminum, which is subject to corrosion attack by mercury. If ered proven and replace the more common shell and tube
the mercury in the feed gas can not be eliminated, then the use design. Demonstrated experience with multiple streams at
of plate fin exchanger systems will provide questionable reli cryogenic temperatures will be necessary before the new
ability. designs will significantly displace the aluminum plate-fin
Thermal Stresses: The problems are very similar to shell 55 exchanger.
and tube exchangers; however, the transient temperature Piping Systems in the Liquefaction Train
analysis is more complicated since the exchanger usually Materials of Construction: In general, natural gas liquefac
transfers heat between two or more process streams. Gener tion can be considered corrosion free since essentially all the
ally, it is advisable to limit necessary temperature variations acid gases are removed in the feed pretreatment area. Thus the
in operation as much as possible. Normally, thermal shock is 60 major concernis the correct use of cryogenic materials. In this
more important in stress evaluation than the normal steady regard, alloy verification is recommended to ensure that what
state temperature difference. is installed is what was specified. Also precautions should be
Leaks: The same considerations as for shell and tube taken to ensure that stress corrosion cracking from chlorides
exchangers apply; however, in a plate fin exchanger, leaks contact (contact with Salt air during construction, glues on
caused by pressure from freezing components such as water, 65 insulation, etc.) does not occur in stainless steel.
oil, etc. are more likely. Thus, keeping the system clean is Flexibility: Due to large temperature variations and result
essential, filters are recommended on the inlet stream to keep ant pipe movement, piping Support and flexibility analysis is
US 8,381544 B2
37 38
paramount. Also, since different materials have different three liquid loading arms are often required depending on
coefficients of expansion/contraction, transition joint analy hydraulics. An additional arm can be used for vapor that is
sis is important when aluminum equipment such as the main displaced and generated in the ship that can be returned back
exchanger or a plate fin exchanger is connected to stainless to the storage tanks to fill the vapor space. The loading lines
steel piping. In the case of long piping runs, an economic can be kept cold during the time then no ship is being loaded
evaluation of pipe loops versus expansion joints should be by circulating a fraction of the plant rundown to the jetty head
performed. and back to the tank. Two designs that are commonly used are
Valve Location: Pneumatic pressure testing of cryogenic two pipelines in the low 20 inch diameter range or one line
systems is preferred so as to avoid water hydrotesting, as about 30 inches in diameter.
residual water can be difficult to remove from valve packings, 10
LNG Tanker Ships (transportation): LNG projects typi
bellows, etc. Pneumatic testing has the disadvantage that a cally require dedicated LNG ships. The number of ships
large amount of energy is stored in the piping during testing. required for an LNG project depends on the distance between
Thus the proper placement of isolation valves reduces this the liquefaction plant and the receiving terminal. LNG trans
stored energy to a manageable level. Additionally, shop
hydrotesting of the piping and a full or near-full non-destruc 15 portation cost increases linearly with distance. The LNG car
tive testing and inspection of field welds further reduces the riers are typically designed for speeds of 17 to 20 knots. The
potential risk. fleet of tankers for an LNG project is a noteworthy portion of
Welds: In an LNG plant, an additional consideration the total cost of the LNG chain. In the LNG ship, the LNG can
regarding welding is the presence of mercury. Aluminum be stored in a refrigerated liquid state while the LNG is
welds (in plate fin exchangerheaders for example) are subject transported. The LNG can be kept cool by evaporating a
to mercury attack. The obvious solution is to eliminate mer fraction of the LNG, which is referred to as boil-off. The ship
cury by pretreatment. If mercury is present, however, the can use the boil-off as fuel for its own engines or can re
following steps help reduce potential damage: set the maxi liquefy the gas. When the ship reaches its destination, the
mum operating temperature at -40 degrees Celsius; avoid LNG can be offloaded to a receiving/unloading terminal. The
severe stress transients on the welds such as during rapid 25 facilities near the receiving/unloading terminal can include
cooldown; avoid mercury accumulation points such as back storage, regasification, and transportation to consumers of
up strips or rings on welds; and ensure that the aluminum natural gas.
system is self draining during a shutdown. Shipping Simulation as Related to the LNG Chain: The
Piping Systems in the Storage and Ship Loading Area most widely used method to date for optimizing the shipping
Transient Analysis: In the storage and ship loading area the 30 system has been an event oriented simulation of the system.
piping runs are typically much longer than in the liquefaction The simulation models the liquefaction plant with appropri
train. A transient analysis should be conducted to avoid exces ate seasonal deviations and maintenance and production con
sive pressure in the long piping system due to the effects of straints. LNG produced in the baseload liquefaction facility is
water hammer produced by emergency shutdown systems, stored in LNG tanks and appropriate losses for heat leaks are
pumps tripping, etc. After a transient analysis and the opti 35 deducted. LNG is loaded from the tanks into the ships, which
mum selection of valve type and valve closure time, the leave the production facility’s harbor, travel to the receiving
system design pressure can be established, which is typically terminal, berth, and prepare to unload. Delays, deviations,
significantly above the normal operating pressure. and boiloff losses for each event along the route are applied.
Bowing Avoidance: If a large cryogenic line is partially After unloading the LNG into the receiving terminals storage
filled with liquid, a significant temperature differential is 40 tanks, the ship, with any residual LNG, sometimes referred to
easily established between the top and bottom of the line as its LNG heel, returns to the liquefaction facility. The ter
resulting in pipe bowing due to greater contraction of the pipe minal vaporizes the stored LNG and exports the vapor to its
bottom compared to the top. This behavior causes excessive consumers based on their demand. Since the Monte Carlo
thermal stresses on the pipe Support system and should be shipping simulation has many random variables, a long time
avoided. Thus the cool down method must be designed to 45 period is simulated to obtain reasonable and reproducible
eliminate uneven top/bottom cooling of the line via the maxi results. For example, the KBR simulation program typically
mum initial use of vapor cool down from the liquefaction uses run lengths of from 5,000 to 10,000 simulated days. In
train. addition, the simulation is usually reinitialized and run sev
LNG Loading eral times to smooth the statistical results. Five to ten of these
The liquefaction facility can also include a ship loading 50 “replicates' are typically run. The usefulness of a simulation
terminal. The LNG can be pumped from a storage tank to an program as an optimization tool depends on the extent in
LNG ship via a loading jetty. At some locations the jetty may which the program meets several general criteria. For simu
need to be extended several kilometers to reach the proper lations, the events that make up a model should closely par
water depth needed, contributing significantly to the plant allel the real system; simplifying assumptions should not
cost. The loading lines can be looped forthermal expansion or 55 adversely influence the simulation results; models should
with expandable bellows. The looped line can have a lower incorporate realistic data and parameters to describe the sys
specific cost, but requires more piping material and jetty tem; numerical results should be accurate, easily quantifiable
space. The bellowed line can be straight thus requiring less and Verifiable; and output should be easy to manipulate and
piping material and jetty space but can also be more expensive analyze.
due to the cost of the bellows. The loading rate is generally 60 Results from Simulations: Some of the major variables that
driven by keeping the ship loading time as short as possible are typically examined by shipping studies include the num
and by the cool downtime of the ship tank material. The LNG ber, size, and speed of ships, the amount of LNG storage and
can be transferred onto the ship through loading arms that both production and receiving facilities, and the number of
have swivel joints, which can allow limited ship movement loading berths. The relationship between these variables and
during LNG transfer before automatically disconnecting. The 65 the total delivered LNG and the cost of shipping can be
height and weight balancing required for easy movement of determined and used as an input into an overall economic
the loading arms can result in very heavy equipment. Two or analysis.
US 8,381544 B2
39 40
Additional Requirements for Shipping Simulations: A from the simulation may be used to analyze the system in
trend of increasingly detailed analysis of shipping systems is detail to find delays in each portion of the system. The results
developing for many projects, which requires more detailed and analyses can be used to design the optimal shipping and
input into the simulations. Specific projects require some or Storage System.
all of the following criteria to be included in their simulations. Cost Reduction in LNG Export Terminals
Other projects find they do not have data available to allow Sophisticated shipping studies can reduce the total LNG
creation of an accurate model or that these details have data storage Volumes required, and thus reduce costs. Certain site
available to allow creation of an accurate model or that these specific cost savings can beachieved, for example by using air
details have little effect on the simulation results. coolers instead of shell and tube seawater coolers. Other cost
In shipping simulations, weather effects in the LNG trans 10 reduction areas can include: (i) use of higher capacity in-tank
portation system, especially for potential bad-weather sites LNG ship loading pumps can enable the use of fewer pumps
and shipping routes. Seasonal variations, latitude and specific in the LNG storage tanks with resultant cost savings; (ii)
geographical area weather variations, and effect on overall optimization of acid gas removal by use of higher concentra
traveling speed and harbor availability also need consider tion solvent with reduced circulation rates e.g. activated
ation. 15 MDEA; (iii) execution of Value Engineering Studies, which
In detailed harbor simulations, harbor geometry and opera can involve a structured exercise including a brainstorming
tions may have an influence on ship operation. Interest is session followed by technical and economic evaluation of
increasing in more detailed models of harbors. Criteria selected ideas; and/or (iv) careful selection of site location to
include specific daylight and tide restrictions, restrictions on avoid excessive site preparation, harbor development, dredg
multiple ship movements, berth and jetty geometry, and load ing, jetty length, etc.
ing rates as a function of number of berthed ships. The installation of single or double containment LNG
In berthing criteria for ships, decisions can be made daily tanks compared with full containment tanks should be con
on which ships should dock first, which berth they should sidered, where there is sufficient plot space and where neces
dock at, and when they should move and load. Simulations sary safety criteria are satisfied.
should model the decisions of human schedulers. Criteria can 25 The use of aero-derivative gas turbine drives for refrigera
include cargo capacity of ships in port, length and cost of tion compressors, e.g., a GE LM6000, has possibilities for
demurrage for each ship, which trades are being served by the cost reduction. These machines are generally lighter and
ships, ship order priority, loading period, next open berth for Smaller than industrial gas turbines, but are more expensive.
loading, and ship and berth restrictions. Higher efficiency and improved reliability can result in
The developmentofan LNG spot market has resulted in the 30 reduced life cycle costs under certain conditions. With
occurrence of short-term trades, with Australia to Spain deliv increased efficiency there is a reduction in fuel gas demand
eries being an example. These trades are inherently difficult to that may require some innovative thinking in the development
model, as they involve market prediction and economic opti of the overall fuel gas balance for the LNG plant.
mization based on a number of constraints due to limited The use of modularization can transfer site construction
resources. A few criteria include economic impact of multiple 35 work to fabrication workshops. Improved productivity can
liquefaction plants serving one trade, LNG price Vs. trade reduce labor costs; however, additional costs are typically
distance (ship utilization) formulas, different pricing levels in incurred due to the increase in Steelwork necessary for trans
different geographic areas, willingness of purchasers to pay portation of the modules. In the extreme, the LNG plant could
premium prices for spot cargos, and contractual constraints. be erected on a purpose built concrete or steel barge which
In simulations of difficulties in ship Scheduling, short-term 40 could be towed to the site and floated into a prepared dock,
scheduling problems, including routing and order Schedul settled down by ballasting and filled in to form a permanent
ing, should be adequately modeled by the logic of the simu foundation. The main advantage of modularization and
lation program itself. Criteria for order Scheduling include barge-mounted facilities is the reduction in site work and site
which terminal ordered first, which terminal needs the LNG labor, which is particularly important in harsh environments
more urgently, whether either terminal is behind on contrac 45 Such as northern Norway, Sakhalin Island and Alaska.
tual deliveries, when the next ship arrives at the liquefaction Design practices and engineering specifications can impact
plant, distances to receiving terminals, and price of LNG at every aspect of the plant, and thus represent an opportunity
each terminal. for significant cost savings. Cost saving exercises often focus
The main point in outlining some of the decisions that the on equipment because the potential cost savings in equipment
simulation logic must handle is not only to demonstrate the 50 are relatively easy to identify. The cost savings related to bulk
difficulty, but also show that the entire system must be treated materials could also be important. Within the process trains
as a whole, rather than as a group of unrelated pieces. The the bulk materials have a cost similar to the equipment. No
actions of one particular ship can be dependent on what each single item is likely to save a large percentage of costs, but
of the other ships and terminals are doing at any particular taken together many Small savings can add up to a sizable
moment. Without taking these short-term scheduling prob 55 total.
lems and interactions into account, the long term averages of Any means of reducing schedule generally has a corre
the simulation will not reflect real fleet operation. LNG ship sponding cost benefit. Interest on capital is reduced and cash
ping and storage systems are an important portion of the flow for the project is improved by earlier sales of LNG. A
overall project economics. Optimizing the shipping and stor number of suggestions have been made for reducing schedule
age is critical to maximizing the return on the investment in 60 Such as: optimization of contract strategy, early placement of
the overall project. Creating a shipping model to simulate the long lead items, optimum site selection, and utilizing an inte
system is an important tool for this optimization. grated project team.
Simulation results can determine the level of LNG delivery Natural Gas Specification
as a function of number of ships, speed of ships, storage Natural gas contains methane, heavier hydrocarbons, and
volume, number of berths, and many other variables. Sensi 65 inert components which all can effect burner performance.
tivity studies can determine the utility of an incremental For this reason, pipeline companies and LNG buyers specify
increase in any of these variables. Extensive data gathered allowable ranges of components and heating values. These
US 8,381544 B2
41 42
requirements can vary widely depending on the market loca ane increases the boil off gas that must be compressed to high
tion. Historically, plant designs have been based on long term enough pressure to be used as fuel or feed to the natural gas
contracts to a limited number of buyers at defined gas speci section of the liquefaction process.
fications, and there was little need for flexibility in the plant Simply injecting -40°C. LPG into the LNG is equivalent
designs, either on the liquefaction or receiving ends of the to pure component refrigeration with methane at -160°C. in
trade. However, the situation is changing as LNG trade a single stage, which is not very efficient. A better approach is
becomes more global. The owners of liquefaction plants can to stage the injection into liquid methane at several pressure
now target more than one market, and new markets may have levels, or use refrigeration to chill the LPG to temperatures
gas specification requirements that are not always compatible closer to LNG temperatures prior to injection. This can be
with existing trades. Furthermore the growing spot market for 10 accomplished by adding an LPG injection pass to the MCHE,
LNG provides opportunities for buyers and sellers who have or adding another exchanger in parallel to the MCHE, using
the ability to be flexible on product specifications. As a result liquefaction level refrigerant to provide the necessary cooling
there is now a desire for technical Solutions regarding condi duty.
tioning of LNG, such as the ability to modify the heating At some facilities there may be a desire to market multiple
value of a product stream at a liquefaction facility. 15 products, such as a high HHV product and a low HHV prod
To prevent liquid dropout, natural gas pipeline companies uct. For a liquefaction plant with a lean feed gas this may be
generally limit the amount of butane, pentane and heavier accomplished by importing LPG and injecting the LPG on a
components that can be in a product. LNG plants must cargo by cargo basis. The LPG injection equipment may be
remove heavier hydrocarbon components to prevent freezing utilized for a cargo destined for a high HHV market, but may
in the liquefaction process, and the heavies removed typically remain idle when loading a ship headed for a low HHV
become a natural gasoline by-product stream. The require market. In theory it is possible to chill the LPG prior to
ments for heating value and gas interchangeability can vary injection using liquefaction refrigeration as mentioned above,
depending on the geographic location of the particular mar but every loading may significantly change operations within
ket. LNG products from worldwide sources can also vary the train. For this reason it may be more efficient to have a
significantly in composition and heating value. 25 stand-alone refrigeration unit, or to rely on greater boil off gas
Early LNG trade was primarily to Japan from Pacific Rim compressor capacity.
and Middle East export plants and to Europe from Northern An important technical feature of the design is to avoid
Africa plants. The Japanese specifications can vary depend sudden vaporization of the LNG as the LPG is injected. This
ing on the importing utility company, but typically have a high can be achieved by pumping the LNG to higher pressure
heating value between 39.7 and 43.3 MJ/Sm (Megajoules 30 before using the LNG to chill the LPG.
per Standard meter cubed at 1 atm and 15°C., which converts If the opposite situation exists where a plant with signifi
to 1065 to 1160 Btu/SCF for a Standard cubic foot at 14.73 cant LPG in the feed is considered for both low and high HHV
psia and 60°F). This relatively high range permits maximum markets, the LPG can be extracted and stored or exported
use of infrastructure by moving greater combustion heat until a high HHV cargo is needed.
capacity for a given Volume. European countries typically 35 One way to accommodate multiple products is based on a
allow wider ranges. Spain, for example, allows a range large facility where two separate products can be produced in
between 35.0 and 44.9 MJ/Smi (940 and 1205 Btu/SCF). different trains with the two products stored separately. This
In one example a gas with HHV-42.6 MJ/Smi is suitable becomes more convenient if two berths are also present, each
for the Japanese and Korean markets, but is too high for the with their ownloading lines. The disadvantage of this method
US or UK markets. In the second example a gas with 40 is the extra storage capacity that would have to be installed
HHV=37.2 MJ/Smi meets US/UK specifications but has a compared to a single product facility.
HHV too low for Korea or Japan markets. Both examples A plant producing multiple products will cost more than a
however are within the ranges allowed for France and Spain. plant marketing a single product. The multiple product plant
Modifying heating value at the liquefaction end usually may still be competitive if the plant is closer to the consumer
means adding or extracting ethane, propane and butane 45 or has economy of scale advantages over single product
(LPG), though nitrogen may also play a part. For natural gas plants in the same market.
Supplies rich in LPG components such as in the Atlantic FIG. 19 depicts a schematic illustration of an embodiment
Basin, a lower high heating value (HHV) is preferred if the of the present invention wherein a cooled stream of LPG is
US and UK markets are to be the consumers. On the other added to an LNG stream to increase the heating value. In FIG.
hand, Pacific Rim consumers prefer a gas with increased 50 19, an embodiment is depicted wherein a cooled stream of
HHV, and Pacific Rim sources that are lean in LPG compo LPG is added to an LNG stream to increase the heating value.
nents may require upward HHV adjustment. Natural gas LNG is used as a cold energy source and is pressurized so that
sourced in the Middle East can physically be shipped to either the LNG that is warmed remains in a liquid state. In process
the Pacific or Atlantic markets, which raises the possibility of 900, the LNG is stored in a storage vessel 910. LNG is
producing two product qualities of differing heating values. 55 pumped out of the storage vessel 910 into line 912 and can be
If LPG must be purchased and injected at the liquefaction delivered for loading onto a transport vessel (not shown). A
location the first challenge is finding a local Source with side stream 914 can be further compressed in pump 916 to
Sufficient quality and then installing the facilities for unload form a higher pressure LNG stream 918 that flows through
ing and storage. More refrigeration is required because the one or more heat exchangers 920 to transfer coldenergy to the
LPG must be chilled from -40°C. (in the case of propane) 60 other stream. The warmed LNG exits heat exchanger920 via
down to -160°C. This requires energy which can be applied line 922, and then flows through a liquid expander 924 to chill
as additional boil off gas compression or refrigeration com and reduce the pressure in line 926 prior to returning to the
pression. If the LPG is injected at -40°C. into the LNG then LNG storage vessel 910. The pump 916 raises the LNG to a
the LPG chilling is accomplished by vaporizing methane, level to ensure that the LNG remains in liquid state after the
which can lead to cavitation within the piping at or near where 65 warming that occurs in the heat exchanger920.
the injection occurs if the injection is done within the process. LPG is stored in a storage vessel 940. LPG is pumped out
If the addition is make in the storage vessels, vaporized meth of the storage vessel 940 into line 942 and can be further
US 8,381544 B2
43 44
compressed in pump 944 to form a higher pressure LPG providing a first liquefied hydrocarbon composed of at
stream 946 that enters the heat exchanger920 to be cooled by least 90 wt % methane, having a temperature of about
cross-exchange with the LNG of line 918. The cooled LPG -150° C. or less and having a first heating value, in a first
exits the exchanger 920 in line 948 and can flow through storage vessel;
control valve 950 prior to blending with the LNG stream 912 providing a second liquefied hydrocarbon composed pri
to form an LNG stream 952 having a higher heating value marily of one or more of ethane, propane, and butane, or
than the LNG prior to blending. mixtures thereof;
A portion of the chilled LPG948 can be diverted to a side providing a first stream of the first liquefied hydrocarbon
stream 954 that can be controlled by valve 956 and returned 10
from the first storage vessel;
into line 942. This recycle loop of line 954 can be used to diverting a first portion of the first stream and compressing
reduce the temperature of the LPG flows 946,948 to tighten it to form a compressed first stream;
the temperature approach of exchanger920 and chill the LPG passing the compressed first stream through the cold side
being added into the LNG to reduce flashing or cavitation of a first heat exchanger to obtain a heated first stream;
upon mixing of the higher temperature LPG into the lower 15 decompressing the heated first stream in a first liquid
temperature LNG. expander to obtain a decompressed first stream;
In one embodiment, the LNG can be stored at a tempera returning the decompressed first stream back to the first
ture of about -160° C. and the LPG can be stored at about storage vessel;
-40°C. The LNG 918 can enter the exchanger920 at about providing a second stream of the second liquefied hydro
-160° C. and exit the exchanger920 at a warmer temperature, carbon;
for example, between -140° C. to about -100° C. The compressing the second stream to form a compressed sec
warmed LNG 922 is then expanded and cooled in the liquid ond stream having a temperature from about -40°C. to
expander 924 prior to reentering the LNG storage 910. The about -160° C.;
LPG can be stored at a temperature of about -40°C. The LPG passing the compressed second stream having a tempera
942 will be at about -40°C. prior to mixing with the chilled 25 ture from about -40°C. to about -160° C. through the
LPG from line 954. The chilled LPG from line 948 can warm side of the first heat exchanger to cool the com
achieve temperatures approaching the LNG inlet temperature pressed second stream to obtain a cooled second stream;
in line 918, for example between -120° C. to about -160° C. combining a first portion of the cooled second stream with
The mixed LPG stream 946 can have a temperature within the 30
a second portion of the first stream to obtain an third
range of between -40°C. to -160° C. depending on the ratio stream having a second heating value greater than the
offlows in lines 942 and 954, generally the mixed LPG stream first heating value of the first liquefied hydrocarbon.
946 will have a temperature within the range of about-80°C. 2. The method of claim 1, wherein the first liquid expander
to about -120° C. provides static expansion to the heated first stream to obtain
The cold approach temperature between the cold LNG 35 the decompressed first stream.
stream 918 entering the exchanger 920 and the cold LPG 3. The method of claim 1, further comprising:
stream exiting 948 is desirably less than 50° C. In alternate diverting a second portion of the cooled second stream and
embodiments the cold approach temperature is less than 35° decompressing it in a second liquid expander to form a
C., less than 25°C., less than 15°C., less than 10°C., or less second decompressed second stream; and
than 59 C. 40 injecting the second decompressed second stream into the
Certain embodiments and features have been described second stream.
using a set of numerical upper limits and a set of numerical 4. The method of claim 3, wherein the second liquid
lower limits. It should be appreciated that ranges from any expander provides static expansion to a portion of the cooled
lower limit to any upper limit are contemplated unless other second stream to obtain the second decompressed second
wise indicated. Certain lower limits, upper limits and ranges 45 Stream.
appear in one or more claims below. All numerical values are 5. The method of claim 1, wherein the flow of the cooled
“about' or “approximately the indicated value, and take into second stream into the first stream is regulated to control the
account experimental error and variations that would be heating value of the third stream.
expected by a person having ordinary skill in the art. 6. The method of claim 1, wherein the temperature of the
Various terms have been defined above. To the extent a 50 cooled second stream is no more than 50° C. warmer than the
term used in a claim is not defined above, the term should be temperature of the first stream.
given the broadest definition persons in the pertinent art have 7. The method of claim 1, wherein the high heating value of
given that term as reflected in at least one printed publication the third stream is 1050 Btu/SCF or greater.
or issued patent. Furthermore, all patents, test procedures, 8. The method of claim 1, wherein the pressure of the
and other documents cited in this application are fully incor 55 heated first stream is sufficient to keep the heated first stream
porated by reference to the extent such disclosure is not in a liquid state.
inconsistent with this application and for all jurisdictions in 9. The method of claim 1, wherein the temperature of the
which Such incorporation is permitted. heated first stream is no warmer than -100° C.
While the foregoing is directed to embodiments of the 10. The method of claim 1, wherein the temperature of the
present invention, other and further embodiments of the 60 cooled second stream is cold enough to prevent cavitation as
invention may be devised without departing from the basic the cooled second stream is combined with the first stream.
scope thereof, and the scope thereof is determined by the 11. The method of claim 1, wherein temperature approach
claims that follow. on the heat exchanger between the heated first stream and the
compressed second stream is less than 20°C.
What is claimed is: 65 12. The method of claim 1, wherein temperature approach
1. A method of altering the heating value of a liquefied on the first heat exchanger between the compressed first
natural gas stream comprising: stream and the cooled second stream is less than 20°C.
US 8,381544 B2
45 46
13. The method of claim 1, further comprising: passing a stream of the LPG having a temperature from
diverting a second portion of the cooled second stream and about -40°C. to about -160° C. through the first heat
decompressing it in a second liquid expander to form a exchanger to obtain a cooled stream of the LPG:
second decompressed second stream, wherein flow of providing a second stream of the LNG;
the second decompressed second stream is regulated to 5 blending a first portion of the cooled stream of the LPG
maintain a temperature approach on the first heat with the second stream of the LNG to obtain a third
exchanger between the heated first stream and the com stream containing LNG and containing increased con
pressed second stream of less than 20°C. tent of LPG and thus having a heating value greater than
14. The method of claim 1, further comprising: the first heating value of the LNG.
diverting a second portion of the cooled second stream and 10 18. The method of claim 17, further comprising returning
decompressing it in a second grid expander to form a the warmed first stream of the LNG exiting the first heat
second decompressed second stream, wherein flow of exchanger to the first storage vessel.
the ninth stream is regulated to maintain a temperature
approach on the first heat exchanger between the com 19. The method of claim 17, wherein the temperature of the
pressed first stream and the cooled second stream at no cooled stream of the LPG exiting the first heat exchanger is no
more than 20° C. 15 more than 50° C. warmer than the temperature of the first
15. The method of claim 1, wherein the temperature of the stream of the LNG entering the first heat exchanger.
second stream is 0° C. or less. 20. The method of claim 17, further comprising blending a
16. The method of claim 1, wherein the temperature of the second portion of the cooled stream of the LPG exiting the
second stream is -30°C. or less. first heat exchanger into the stream of the LPG prior to enter
17. A method of modifying the heating value of a liquefied ing the first heat exchanger.
natural gas stream comprising: 21. The method of claim 1, wherein the compressed second
providing LNG composed of at least 90 wt % methane, and stream, has a temperature from about -80°C. to about -120°
having a first heating value in a first storage vessel; C.
providing LPG composed primarily of one or more of 22. The method of claim 17, wherein the stream of the LPG
ethane, propane, and butane, or mixtures thereof; 25 has a temperature from about -80° C. to about -120° C.
passing a first stream of the LNG through a first heat
exchanger to provide cooling energy and a warmed first
stream of the LNG;

United States Patent (10) Patent No.: US 8,381544 B2: Coyle (45) Date of Patent: Feb. 26, 2013

Uploaded by

Copyright:

Available Formats

United States Patent (10) Patent No.: US 8,381544 B2: Coyle (45) Date of Patent: Feb. 26, 2013

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

United States Patent (10) Patent No.: US 8,381544 B2: Coyle (45) Date of Patent: Feb. 26, 2013

Uploaded by

Copyright:

Available Formats

USOO8381544B2

(12) United States Patent (10) Patent No.: US 8,381544 B2

You might also like