Safety History of International LNG Operations
Safety History of International LNG Operations
of
International LNG Operations
January 2005
H
|
H —C— H
|
H
LNG
LNG
Prepared by ~
H
H C H
H
CH·IV International
The LNG Specialists
1120C Benfield Boulevard
Millersville, MD 21108
Phone: 410-729-4255
Web Site: CH-IV.com
Technical Document
TD-02109
H
H C H
H
CH·IV International
INTRODUCTION:
LNG has been safely handled for many y ears. The industry is not without its
incidents and accidents, but it maintains an enviable “modern-day ” 1 safety record.
The process of natural gas liquefaction, storage and vaporization is not a new
technology . Earliest patents involving cryogenic liquids date back into the mid-
1800s. The first patent directly for LNG was awarded in 1914. In 1939, the first
commercial LNG peak-shaving plant was built in West Virginia. There are over 120
peakshaving and LNG storage facilities 2 worldwide, some operating since the mid
1960s. In addition, there are over 18 base-load liquefaction (LNG export) facilities in
Abu Dhabi, Alaska, Algeria, Australia, Brunei, Indonesia, Liby a, Malay sia, Oman,
Nigeria, Qatar and Trinidad currently in operation. LNG is transported by a fleet of
over 130 LNG tankers of vary ing sizes from 18,500 M 3 (cubic meter) to 140,000 M 3 .
This fleet of LNG ships delivers to receiving terminals in the Belgium, Dominican
Republic, France, Greece, Italy , Japan, Korea, Spain, Taiwan, Turkey , the U.K. and,
of course, the U.S., including Puerto Rico.
The LNG storage tanks at these facilities are constructed of an interior cry ogenic
wall, usually made of 9% nickel steel, aluminum or other cry ogenic alloy . The
outside wall is usually made of carbon steel or reinforced concrete. A thick lay er of
an insulating material such as Perlite separates the two walls.
1
Modern Day – Post mid-1950s - Cryogenic technologies came of age during the late 1950s and early 1960s with
the development of the U.S. space program where cryogenic fuels such as liquid hydrogen and liquid oxygen
had to be routinely and safely handled.
2
This does not include dozens of small LNG vehicle fueling stations and industrial LNG fuel facilities.
3
There has never been a catastrophic tank failure with any LNG, or similarly designed, storage tank.
plant personnel. Other accidents have occurred during the construction and repair of
LNG facilities. Some of these accidents have been used to tarnish the exceptional
safety record of LNG, but as no LNG was directly involved in the incident these
accidents can only truly be called “construction” accidents. Damage has alway s been
limited to the plant proper.
The following three sections discuss land-based, LNG ship and over-the-road LNG
transport incidents respectively . Each section references an appendix listing the
various incidents.
S A F E T Y R E C O R D O F L A N D -B A S E D LNG F A C I L I T I E S
The first commercial facility for producing or utilizing LNG was a peakshaving plant
that began operations in 1941 in Cleveland, Ohio. (A peakshaving plant liquefies
natural gas when customer demand for gas is low and then vaporizes the LNG when
demand is high, thus handling periods of peak demand that cannot be met by existing
gas pipelines.) Since then, more than 150 other peakshaving plants have been
constructed worldwide (approximately one-half of these are satellite facilities that
have no liquefaction capability ). In addition, 18 large natural gas liquefaction plants
(export facilities) and about 30 large LNG import terminals have been constructed.
There have been four incidents in operating LNG facilities directly attributable to
LNG that resulted in one or more fatalities – Skikda, Algeria – 2004; P. T. Badak
(Bontang, Indonesia), 1983; Cove Point Mary land, 1979; Arzew, Algeria, 1977; and
Cleveland, Ohio, 1944. There were two other “LNG” incidents (Portland 1968 and
Staten Island 1973) involving death, but these correctly should be classified as
construction accidents since no LNG was present. See Appendix A for more details
on these incidents and a complete listing of land-based LNG facility incidents.
The accident at East Ohio Gas Company ’s peakshaving plant in Cleveland, Ohio, is
the only incident that involved injuries or fatalities to persons not employ ed by the
LNG facility or by one of its contractors. This accident is often used as an example
of the danger or risk involved in the LNG industry . However, the industry has
changed dramatically since 1944. Modern LNG plants are designed and constructed
in accordance with strict codes and standards that would not have been met by the
Cleveland plant. For example, the alloy used in Cleveland for the inner vessel of the
LNG storage tank is now forbidden and each LNG tank must now be located within a
dike capable of containing at least 100% of the tank’s capacity .
S A F E T Y R E C O R D O F LNG S H I P S
The first transportation of LNG by ship took place early in 1959 when the Methane
Pioneer (an ex-Liberty ship that had been extensively modified) carried 5,000 M 3
(cubic meters) of LNG from Lake Charles, Louisiana, to Canvey Island, near London,
England. Commercial transportation of LNG by ship began in 1964 when LNG was
transported from Arzew, Algeria to Canvey Island in two purpose-built ships—the
Methane Princess and the Methane Progress.
The overall safety record compiled by LNG ships during the thirty -nine y ear period
1964 - 2002 has been remarkably good. During this period, the LNG tank ship fleet
has delivered more than 30,000 shiploads of LNG, and traveled more than 100 million
kilometers while loaded (and a similar distance on ballast voy ages).
In all of these voy ages and associated cargo transfer operations (loading/unloading),
no fatality has ever been recorded for a member of any LNG ship’s crew or member of
the general public as a result of hazardous incidents in which the LNG was involved.
In fact, there is no record of any fire occurring on the deck or in the cargo hold or
cargo tanks of any operating LNG ship.
Among LNG import and export terminal personnel, only one death can be even
remotely linked to the loading or unloading of LNG ships. (In 1977, a worker in the
LNG Export Facility at Arzew was killed during a ship-loading operation when a
large-diameter valve ruptured and the worker was spray ed with LNG. His death was
the result of contact with the very cold LNG liquid; the spilled LNG did not ignite.
See Item 6 in Appendix A.)
O V E R - T H E R O A D LNG T R A N S P O R T A C C I D E N T S
SUMMARY
The various incidents discussed, when taken on a case-by -case basis, attests to LNG’s
safety record. The fact that most LNG opponents cite Cleveland and Staten Island as
examples of the dangers of LNG, clearly indicate that there is little else to make their
point. As devastating as both Cleveland and Staten Island were, they have no
relevance when discussing the design and operation of today ’s LNG facilities.
LNG is cry ogenic; it is a liquid; and its vapors are flammable. It is not without its
safety concerns – it, however, can be produced, transported and revaporized as safely ,
and in most cases, more safely , than other liquid energies.
For more information on LNG safety , please see CH·IV’s website, particularly :
http://www.CH-IV.com/lng/lngsafty.htm
On October 20, 1944, the tanks had been filled to capacity in readiness for
the coming winter months. About 2:15 PM, the cy lindrical tank suddenly
failed releasing all of its contents into the nearby streets and sewers of
Cleveland. The cloud promptly ignited and a fire ensued which engulfed the
nearby tanks, residences and commercial establishments. After about 20
minutes, when the initial fire had nearly died down, the sphere nearest to the
cy lindrical tank toppled over and released its contents. 9,400 gallons of LNG
immediately evaporated and ignited. In all, 128 people were killed and 225
injured. The area directly involved was about three-quarters of a square mile
(475 acres) of which an area of about 30 acres was completely devastated.
The Bureau of Mines investigation showed that the accident was due to the
low temperature embrittlement of the inner shell of the cy lindrical tank. The
inner tank was made of 3.5% nickel steel, a material now known to be
susceptible to brittle fracture at LNG storage temperature (minus 260°F). In
addition, the tanks were located close to a heavily traveled railroad station
and a bombshell stamping plant. Excessive vibration from the railroad
engines and stamping presses probably accelerated crack propagation in the
inner shell. Once the inner shell ruptured, the outer carbon steel wall would
have easily fractured upon contact with LNG. The accident was aggravated
by the absence of adequate diking around the tanks, and the proximity of the
facility to the residential area. The cause of the second release from the
spherical tank was the fact that the legs of the sphere were not insulated
against fire so that they eventually buckled after being exposed to direct
flame contact.
4
“Report on the Investigation of the Fire at the Liquefaction, Storage, and Regasification Plant of the East Ohio
Gas Co., Cleveland, Ohio, October 20, 1944,” U.S. Bureau of Mines, February, 1946.
Further, it should be noted that the ignition of the two unconfined vapor
clouds of LNG in Cleveland did not result in explosions. There was no
evidence of any explosion overpressures after the ignition of the spill from
either the cy lindrical tank or the sphere. The only explosions that took place
in Cleveland were limited to the sewers where LNG ran and vaporized before
the vapor-air mixture ignited in a relatively confined volume. The U.S.
Bureau of Mines, concluded that the concept of liquefy ing and storing LNG
was valid if “proper precautions are observed.”
The Cleveland Disaster put an end to any further LNG development in the
United States for many y ears. It was not until the early sixties that LNG
began to be taken seriously through construction of LNG peakshaving
facilities. A number of elements came together to bring LNG back; these
included:
• The advent of the space program and its associated cry ogenic technologies
• Successful large-scale fire and vapor cloud dispersion demonstrations
• Extensive cry ogenic material compatibility studies
• Construction and operation of liquefaction plants in Algeria and receiving
terminals in France and England.
5
See Section 3.1 of CH·IV’s “Introduction to LNG Safety,” Short Course on LNG Rollover.
The subsequent New York City Fire Department investigation 6 concluded that
the accident was clearly a construction accident and not an LNG accident.
This has not prevented LNG’s opponents from claiming that since there may
have been latent vapors from the heavy components of the LNG that was
stored in the tank, then it was in fact an LNG incident.
6
"Report of Texas Eastern LNG Tank Fatal Fire and Roof Collapse, February 10, 1973," Fire Department of
the City of New York, July, 1973
7
“Columbia LNG Corporation Explosion and Fire; Cove Point, MD; October 6, 1979" National Transportation
Safety Board Report NTSB-PAR-80-2, April 16, 1980
8
A flat plate temporarily installed between flanges during construction and/or maintenance to isolate equipment.
There have been other accidental ignitions involving LNG during large-scale
tests.
• One occurred in England during large-scale fire tests being carried out by
British Gas Corporation. Stray currents from a nearby radar station were
blamed for prematurely igniting the primer that was eventually to be used
to ignite the LNG cloud.
• Another occurred in Japan during similar large-scale tests carried out by
Japan Gas Association. The ignition mechanism was not explained.
• During a test at a research facility near San Clemente, California, a sudden
change in wind direction caused the vapor cloud to encounter a tractor that
was moving some of the test equipment. The tractor ignited the vapor
cloud, badly burning the driver. A researcher was also in the vapor cloud
at the time of ignition. He was able to get out of the vapor cloud before
the flame front reached him by running crosswind and was not injured.
9
See description in Section 3.1 of CH·IV’s “Introduction to LNG Safety”
1. 1964/1965
25,500 M 3 Jules Verne
While loading LNG in Arzew, Algeria, lightning struck the forward vent riser
of the ship and ignited vapor, which was being routinely vented through the
ship venting sy stem. Loading had been stopped when a thunderstorm broke
out near the terminal but the vapor generated by the loading process was being
released to the atmosphere. The shore return piping had not y et been in
operation. The flame was quickly extinguished by purging with nitrogen
through a connection to the riser.
A similar event happened early in 1965 while the vessel was at sea shortly
after leaving Arzew. The fire was again extinguished using the nitrogen purge
connection. In this case, vapor was being vented into the atmosphere during
ship transit, as was the normal practice at that time.
2. May, 1965
27,400 M 3 Methane Princess
The LNG loading arms were disconnected before the liquid lines had been
completely drained, causing LNG to pass through a leaking closed valve and
into a stainless steel drip pan placed underneath the arms. Seawater was
applied to the area. Eventually , a star-shaped fracture appeared in the deck
plating in spite of the application of the seawater.
3. May, 1965
25,500 M 3 Jules Verne
On the fourth loading of Jules Verne at Arzew in May 1965 an LNG spill,
caused by overflowing of Cargo Tank No.1, resulted in the fracture of the
cover plating of the tank and of the adjacent deck plating. The cause of the
overfill has never been adequately explained, but it was associated with the
failure of liquid level instrumentation and unfamiliarity with equipment on the
part of the cargo handling watch officer.
5. September, 1968
5,000 M 3 Aristotle
Ran aground off the coast of Mexico. Bottom damaged. Believed to be in
LPG service when this occurred. No LNG released.
7. September 2, 1970
71,500 M 3 Arctic Tokyo
Sloshing of the LNG heel in No. 1 tank during bad weather caused local
deformation of the primary barrier and supporting insulation boxes. LNG
leaked into the interbarrier space at one location. No LNG released.
8. Late 1971
50,000 M 3 Descartes
A minor fault in the connection between the primary barrier and the tank dome
allowed gas into the interbarrier space. No LNG released.
9. June, 1974
27,400 M 3 Methane Princess
On June 12, 1974 the Methane Princess was rammed by the freighter Tower
Princess while moored at Canvey Island LNG Terminal. Created a 3- foot
gash in the outer hull. No LNG released.
13. 1974
27,400 M 3 Methane Progress
Ran aground at Arzew, Algeria. Damaged rudder. No LNG released.
23. 1985
35,500 M 3 Isabella
LNG released as a result of overfilling a tank. Deck fractured due to low
temperature embrittlement.
24. 1985
35,500 M 3 Annabella
Reported as “pressurized cargo tank.” Presumably , some LNG released from
the tank or piping. No other details are available.
25. 1985
126,000 M 3 Ramdane Abane
Collision while loaded. Port bow affected. No LNG released.
The Motorship “Yuyo Maru No. 10” (gross tonnage of 43,723), laden with 20,831 MT of light
naphtha, 20,202 MT of propane and 6,443 MT of butane, left Ras Tanura, in the Kingdom of Saudi
Arabia, for Kawasaki, and the port of Keihin on October 22, 1974. While the vessel was sailing
northward along the Naka-no Se Traffic Route in Tokyo Bay on November 9, she collided with the
Motorship “Pacific Ares” (gross tonnage of 10,874), manned with a Taiwanese Master and 28 crew
members, laden with 14,835 MT of steel products, en route from Kisarazu for Los Angeles, USA.
The collision occurred about 13:37 hours on the same day slightly northward of the boundary line
of the Naka-no Se Traffic Route.
As a result of the collision, the “Yuyo Maru No. 10” suffered a large hole at the point of collision,
with her cargo naphtha [The naphtha was carried in its outer ballast tank (between the insulated
LPG tanks and the hull of the ship). This is effectively what makes up the “double hull” with LNG
ships. The LPG cargo tank was not penetrated. LNG tankers never carry any thing other than air
or ballast (water) in these tanks.] instantly igniting into flames. As a result of the outflow of
naphtha overboard, the sea surface on her starboard side literally turned into a sea of fire. The
“Pacific Ares” showered with fire burst into flames in the forecastle and on the bridge. While
explosions occurred one after another [naphtha, not propane], attempts were made to tow the
“Yuyo Maru No 10”, outside the bay, but she ran aground in the vicinity of Daini Kaiho. She was
successfully towed out of Tokyo Bay and sunk south of Nojima Saki on the afternoon of November
27 [Thirty-six days after the original collision.] by cannon, air bomb and torpedo attacks staged by
the Maritime Self-Defense Force. [Please note “cannon, air bomb and torpedo attacks” were
required to sink the ship. Other reports indicate that these attacks lasted one and a half days. The
author has seen a black and white film of these attacks. It appeared that the LPG tanks were for
the most part fully in tact prior to the attacks. The ship’s LPG vent stacks were melted down to just
above the decks and on fire indicating that LPG remained within the storage tanks.]
On board the “Yuyo Maru No. 10”, five crew members were killed and seven others injured by this
accident. The “Pacific Ares”, whose forward section was completely crushed and superstructures
burned down, was later repaired. Her crew members were all killed except one person, who was
injured but rescued.
Note: The last four incidents were reported on television and/or presented in
the local Boston print media. In every case the media attempted to
create a disaster scenario using meaningless phases such as “blast
zone” and “police cruisers turned off lights to prevent explosions.” In
one case a totally misinformed fire chief stated that the situation was
“potentially a giant bomb. . . . An explosion would devastate a half-
mile in all directions.” One of the worst “facts” reported was that
“water was hosed onto the tanker to keep the LNG cool”!