Royal Dutch Shell Plc  .com Rotating Header Image

Shell Prelude: Tales of the Unexpected – When the party ended with a bang!

Screen Shot 2014-04-03 at 00.47.46Tales of the Unexpected – When the party ended with a bang!: 3rd in a series of articles by Bill Campbell (right), retired HSE Group Auditor, Shell International, about safety issues relating to the Shell Prelude FLNG project

Screen Shot 2014-04-02 at 18.10.18

Thus methane-air explosions are unpredictable, and by definition unpredictable events take you by surprise and can occur when you least expect them, and often when you are least prepared.  And unfortunately, from time to time, these unpredictable events can have catastrophic consequences as history tells us.

By Bill Campbell

LNG is natural gas (methane) refrigerated, the chilling process eventually turning the gas into a liquid shrinking its volume by 600  times. As we are aware from elementary physics, energy cannot be created or destroyed, so the whole economic model of the use and transportation of LNG worldwide, which really started in the 50’s and is due to exponentially expand in the next decade, is that the heat energy contained in one metre cubed of the liquid equates to six hundred metres cubed of methane.  So the physical characteristics of liquified natural gas is what makes it economically viable in its transportation over in some cases many thousands of miles from its source to where it will be used when converted again into its natural state.  But it’s this conversion that can make it so dangerous should it spill or leak into the atmosphere accidentally.

In the US, the use and transportation of LNG was curtailed and did not take off again till the early 70’s.  It was during the war, and this may have been a factor, but in 1944 an LNG tank ruptured in Cleveland (inadequate nickel content in the stainless steel tank).  The resulting explosion and fire affected 47 surrounding acres, the black and white photographs are reminiscent of pictures of post bomb Hiroshima.  The resulting fire and explosion killed 128 people.

Fortunately, spills of LNG in volumes that can have catastrophic consequence are rare but in recent years are increasing in frequency.  Human error as ever, seems a key causal factor. In 2004 at Skikda, Algeria, an explosion and fire caused 27 deaths, 74 injuries and material damage well outside the plant boundary.  In the same year a pipe carrying LNG was accidentally damaged in Belgium, 23 known fatalities and in 2005 in Nigeria an LNG underground pipe exploded and the resulting fire engulfed an estimated 27square kilometres.

So what?

Well the good news is that many other LNG spills occurred with no serious consequence, it would appear that it is the physics of LNG and how it vaporises back into its gaseous state that saves the day.  When spilled, being heavier than air it sinks until it contacts the ground or the sea.  The dispersion of the vapour cloud then depends on many variables, mainly volume spilled, wind speed and direction, contact with water or seawater etc.  If large volumes are spilled, the expansion ratio can lead to vapour clouds measured in cubic kilometres rather than cubic metres but much of the bubble is either lean, below the flammability limits, or rich, above the flammability limits of methane.  God, or whichever God you believe in, wisely made methane a gas with a restricted flammability range, from circa 5 to 15 per cent by volume in air.  Nevertheless it is still the major industrial killer, and causes significant deaths domestically as well.

Concerned about the potential consequences of LNG spillage, accidental or due to terrorist activities,much research was carried out worldwide. Shell also involved itself in trials with LNG spillage at sea, videos were taken but restricted in distribution.  The bottom line was that LNG spills were difficult, or nigh impossible to ignite. Within the large and expanding vapour cloud you need to have a subset bubble within the flammable range to come into contact with a source of ignition.

Getting back to our vested interest in a Prelude, and offshore installations in general.

There are many hundreds of significant gas releases with no consequential damage. From UK North Sea official statistics over the two year period 2008/9 for example, there were 155 gas leaks of significance or one every 5 days or so.  This data covers some 200 installations. So the probability of gas leaks on offshore installations is ever present, and historic data means that in the formulation of an offshore installation risk matrix for Prelude FLNG say, such a leak, whether they be raw gas, gas condensate, LPG, or LNG must be considered as a likely if infrequent event. The 155 North Sea leaks used in the example above, defined in The UK legislation as dangerous occurrences because of the potential of these events, turned out to be of little consequence  as no ignition of the flammable atmosphere occurred or if it did ignite there was insufficient fuel to cause a damaging explosion

So in summary, explosions of methane air mixtures appear to be randomly distributed, that is the probability of an explosion today, is the same as it was yesterday, and the same as it will be tomorrow. If this gives you false hope it should not. The fatal explosion on Cormorant Alpha in 1982 was the first time gas had entered a non hazardous area on this facility, the explosion on the same facility in 1989 was the first time gas in any quantity had been released in a concrete column, in 1988 the explosion on Piper A was preceded by a significant gas release a year or so earlier which did not ignite, the explosion on Deepwater Horizon appears to have occurred after only two previous occasions when drill kicks caused significant volumes of gas in the atmosphere, and the minor ignition of gas within the accommodation on Fulmar was the first known time that gas had entered such a building.

Thus methane-air explosions are unpredictable, and by definition unpredictable events take you by surprise and can occur when you least expect them, and often when you are least prepared.  And unfortunately, from time to time, these unpredictable events can have catastrophic consequences as history tells us.

This can be best demonstrated in the short but true tale below.

The party was a dull affair until unexpectedly things went off with a bang

This is a true story, a story of how unpredictable but constantly dangerous LNG spillages can be. At a safe and well designed test facility located far from the madding crowd, at Mercury in the Nevada Desert, in 1987, the US Dept of Energy huffed and puffed, day after day, doing large scale tests to get LNG spills to ignite. Frustrated, no doubt with their lack of success the boffins were ready to pack bags to go home when a vapour cloud accidentally ignited and damaged and propelled polyurethane pipe and insulation outside the test site boundary fence.  It is not known or reported that anyone was injured, so hopefully the only cost was an increased laundry bill.

Bill Campbell

Comment posted on Shell Blog by “Cash All Gone” on Apr 3rd, 2014

Reaction to Bill Campbell’s scaremongering article – please get your facts straight as the article is a brilliant mix of truth, halftruth and false statements. For example – the Cleveland incident in 1944 – there was no LNG explosion, there was fire only, except minor explosions in the sewage where the LNG got confined – source:

The explosion in Belgium was not an LNG pipeline, the only regasification facility in Belgium is in Zeebrugge, at the offloading terminal. The gas explosion was caused by a bulldozer hitting a high-pressure maim ring pipeline running under an industrial area. Nothing to do with LNG except that the gas in the line may have been supplied into Belgium as LNG, but it was in gaseous form, nit liquid form in that pipeline.

I see the reasons for your gripe with Shell, but please stay with facts, not with assumptions. You have left Shell and the industry a long time ago – the industry has changed and is still changing rapidly, especially on the HSE side, as it is well understood that it is an essential part of the license to operate.

Comment posted by Outsider on Apr 3rd, 2014 at 17:33

Cash All Gone: I think you are overdoing it a bit. Bill describes a number of incidents involving methane. LNG is liquid methane, but in the event of release it will immediately assume the gaseous form, which when mixed with air can form explosive mixtures. You are right in that LNG in liquid form is not explosive, but the time for which it remains liquid in the event of release is very short. Any escape of LNG will therefore result in a potentially explosive cloud of methane gas and air – exactly as happened in the events that Bill describes.

BILL CAMPBELL RELY. Posted on 2014/04/03 at 19:18

Cash all gone comments – My reply


Could you please include this correction as a blog from me. if the fellow cares to read the report on the website he linked it clearly states that when the fire had almost died down, the tank collapsed with some 9400 gallons of LNG spilled, this evaporated and in their words promptly ignited so we had a fire followed 20 minutes later by an explosion.

The Belgium incident is listed as an LNG incident on many websites.

I don’t quite understand the scaremongering charge, after all, these events happened did they not and were generally as a result of human error. It is to avoid the charge of bias that at all times I refer to historical failure data both related to Shell operations but industry operations also, as Cash be Gone should understand in determination of the probability of such events were are guided by – has it happened in our industry, has it happened to Shell, the answer to many of the questions is yes it has.



Comment received by email from “Relieved” on Apr 3rd 2014

Here are a couple of articles about the effects of natural gas explosions.

Gas Explosion Tears Through Texas Pastures – New York Times: Apr 8, 1992 – … an explosion that apparently resulted from a leak of a large amount of liquefied petroleum gas from a pipeline outside this east central Texas …

Explosion in Rural Hamlet Raises Troubling Questions – New York …: The New York Times: Published: March 16, 1990 … The blast, which blackened a swath of farm land and trees, killed …

With regards to the first article, this explosion shock my house and rattled windows and I was located over 100 miles from the detonation site, i.e., this explosion shook up homes on the outskirts of Houston, TX. The local media went crazy for about a week because of this incident.

If there should be any leak of significance on the Prelude a gas explosion would turn that vessel into scrap iron.

Posting on Shell Blog by “Cash all Gone” on 2014/04/04 at 14:02

To Bill: if you read the report I refer to, it says that the contents ignited. It even says in the last paragraph:


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 cylindrical 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 liquefying and storing LNG was valid if “proper precautions are observed”.


Regards the accident in Belgium, please google a bit, the accident was a ruptured high pressure gas line (40in, 85bar). Please check the place – it is more than 100km away from the regasification facility in Zeebrugge.

Regarding scaremongering – maybe I overreacted a bit, you can never exclude risk fully, but it is not exclusive to Shell and should not be an excuse to not try to evolve technology to the next level… I may be mistaken on the purpose of your article, but it seems that that is your message.


Related articles

This website and sisters,,,, and, are owned by John Donovan. There is also a Wikipedia segment.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.