Report sheds new light on LNG blast in Algeria

Apr 14, 2004 02:00 AM

A newly released document provides important insights into the chain of events that led to the January explosion of a LNG facility in the African nation of Algeria. Several scientists who specialize in LNG research said the document indicates that a similar accident could occur at LNG plants like those proposed for Mobile Bay and elsewhere in the United States.
Initial reports blamed a faulty steam boiler for the massive explosion and fire at the government-owned Skikda, Algeria, plant. Those reports were incorrect, according to the new document presented by Sonatrach, owner of the destroyed LNG plant. A display titled "The Incident at the Skikda Plant: Description and Preliminary Conclusions" indicates, instead, that a large amount of liquid gas escaped from a pipe and formed a cloud of highly flammable and explosive vapour that hovered over the facility. The cloud exploded after coming into contact with a flame source.

The exact nature of the cloud is likely to be sharply debated as industry advocates and even a number of independent scientists have argued that an LNG vapour cloud, if it were to form, would be relatively small and would not explode. Most of the 27 people who died were killed by the force of the blast, according to the report. The report lists a "few casualties by fire," though the fire burned for eight hours.
The Sonatrach report was presented at an international LNG conference held in the Middle Eastern nation of Qatar in late March. Officials with the US Department of Energy (DOE), the Federal Energy Regulatory Commission (FERC) and ExxonMobil declined to discuss the document with the Mobile Register.

In the days after the accident, officials with the DOE, FERC and ExxonMobil, as well as Alabama Port Authority director Jimmy Lyons, stressed that the explosion seemed to be entirely related to a malfunctioning boiler. LNG plants in the United States, they argued, would not have boilers like the ones used at the plant in Algeria, so a similar accident could not occur at an LNG facility in America.
But several scientists who examined the new report told the Mobile Register that the type of accident described in it could occur at an LNG facility in this country, regardless of the type or number of boilers present. Almost any source of ignition, from a cigarette lighter to a pilot light, could have ignited a vapour cloud.

ExxonMobil and Cheniere Energy have both proposed building LNG facilities on the shores of Mobile Bay, close to residential neighbourhoods. Both companies said their facilities would not impact nearby residents, even in the event of a catastrophic accident. ExxonMobil would place its plant on land owned by the Port Authority at the former Navy home port; Cheniere would build on Pinto Island.
"I think this tells us that dealing with LNG is a tricky and dangerous business," said James Fay, professor emeritus at the Massachusetts Institute of Technology and one of the nation's leading LNG scientists. "It was apparently a very large gas leak that went on for a while before the explosion. That certainly doesn't give you a lot of faith in their gas detection equipment, with all this gas leaking out. I guess this means sometimes that equipment doesn't work."

Fay said the failure may have important implications for the siting criteria used by FERC when granting permits for new onshore LNG facilities. In particular, Fay said, FERC requires only that companies prove they can contain a vapour cloud and fire resulting from a 10-minute leak of LNG at the plant.
"The fire burned for eight hours, and that fact does seem unusual. I would have thought it would have burned up more quickly," Fay said. "Maybe there wasn't anyone to shut the equipment down. Maybe all of the workers perished in the blast, and the equipment just kept running, spewing LNG out so it just kept burning and burning. ... FERC's rules just say a company would have a 10-minute leak. That's it. But clearly this one kept leaking for a much longer time period."

Fay and others said the report is missing a critical piece of information: Whether the fuel that leaked from the pipe at the plant was LNG or a LPG, such as propane, or some combination of both. LNG and LPG were present in some quantities at the Skikda plant, the report said, though the damage to the facility was so extensive, it may be impossible to know exactly what kind of gas formed the vapour cloud.
Few would be surprised if LPG proved to be the culprit -- the vapours are known to be highly volatile, and prone to explode when exposed to flame. Pure LNG -- which is almost 100 % methane -- usually is thought to explode only in confined spaces, such as a building or the hull of a ship, according to scientists.

In presentations made in Mobile by the DOE, FERC and ExxonMobil, officials stressed that "LNG does not explode." They also said that if an LNG vapour cloud formed and was somehow ignited, the flame would move through the cloud so slowly that a person simply could walk ahead of it and stay out of danger.
While some scientists agree that may be true of "pure" LNG, which would be entirely methane, the scientific literature suggests that much of the LNG shipped to facilities around the country typically is contaminated with some quantity of more explosive "LPG" gases, such as propane.

A 1980 Coast Guard study titled "LNG Research at China Lake," states that LNG imported into this country is often far from pure, and it reveals that vapour clouds made from "impure" LNG actually explode as readily as the highly volatile LPG. When natural gas is super-cooled and turned into a liquid, as much as 14 % of the total cargo shipped as LNG may actually be LPG or other hydrocarbon fuels, according to the Coast Guard report. Natural gas contains these other fuels when it is pumped from the ground.
LNG containing these so-called "higher hydrocarbons" is known as "hot gas" and has a higher energy content than pure methane. The Coast Guard report reveals that vapour clouds of LNG containing at least 13.6 % of these other fuelscan detonate just like pure propane gas. The agency concluded in its report that this deserves "special consideration, as the commercial LNG being imported into the US East Coast has about 14 % higher hydrocarbons."

Several scientists said they were unaware of the Coast Guard's report. They also were unaware that LNG arriving in the United States sometimes contained significant quantities of other gases, such as propane, butane and ethane. They agreed that in light of the Skikda incident, statements made by the LNG industry and federal officials regarding the explosive potential of LNG vapour clouds may need to be re-examined.
"It's pretty clear that this was not sabotage," Fay said, discounting rumours that terrorists may have tried to damage the facility. "I think there is a strong suspicion that the explosion which occurred could have been an LPG explosion or an LNG explosion. If it were LNG, this would be the first major LNG explosion that occurred anywhere." It is also one of the largest vapour cloud explosions on record, according to scientists.

"The fact that there was a vapour cloud is huge," said Bill Powers, an engineer based in California who has studied LNG terminals, siting issues for both onshore and offshore proposals. "We don't know if it was an LNG vapour cloud or an LPG cloud or a mix of both, but, either way, it means it is the kind of accident that could happen here."
Powers pointed out that several terminals proposed for the United States would deal with both LPG and LNG. At the terminal proposed for Long Beach, California, for instance, Powers said the LPG tanks would be right next to the LNG facility. Powers also felt it was noteworthy that Halliburton had conducted a major renovation of the Skikda plant in 1999, updating all of the key safety equipment and computer systems.

A Halliburton website touts the revamped LNG terminal as a model of modern American workmanship.
"Halliburton is pleased to announce that its recently completed LNG Revamp Project at Skikda, Algeria, has passed all its performance tests," reads the company news release announcing the project's completion. "KBR's work included extensive revamp of the three LNG trains and associated utilities and auxiliaries and a complete revamp of the complex's electrical power and control systems. ... Over 9,000,000 construction man-hours were expended."
The three separate LNG regasification plants or "trains" that were revamped by Halliburton were destroyed in the explosion.

Powers said Halliburton's engineers had missed a weak link in their safety planning for the facility.
"That highlights the importance of putting these facilities in places where, no matter what, people will not be at risk. If a company like Halliburton missed a scenario that could cause this, that tells us that we cannot account for all possible accident scenarios at LNG facilities," Powers said.
"Halliburton would have exhaustively checked out every possible accident chain of events and accounted for it, countered it," he said. "They would do that before they give it a clean bill of health. That's how they operate. They must have simply missed this accident possibility."

Source: Washington Times
Market Research

The International Affairs Institute (IAI) and OCP Policy Center recently launched a new book: The Future of Natural Gas. Markets and Geopolitics.


The book is an in-depth analysis of some of the fastest moving gas markets, attempting to define the trends of a resource that will have a decisive role in shaping the global economy and modelling the geopolitical dynamics in the next decades.

Some of the top scholars in the energy sector have contributed to this volume such as Gonzalo Escribano, Director Energy and Climate Change Programme, Elcano Royal Institute, Madrid, Coby van der Linde, Director Clingendael International Energy Programme, The Hague and Houda Ben Jannet Allal, General Director Observatoire Méditerranéen de l’Energie (OME), Paris.

For only €32.50 you have your own copy of The Future of Natural Gas. Markets and Geopolitics. Click here to order now!


Upcoming Conferences
« June 2018 »
1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30

Register to announce Your Event

View All Events