A very large area affected by a low pressure wave, less than 2 psi, rather than 'detonation' is the more technically correct term when describing the explosion that devastated the Buncefield Oil Storage Terminal and the surrounding community near London in December 2005. However, the distinction is largely academic, said fire protection specialist David White.
"The difference between the two is just a few thousand feet per second," said White. "With the massive damage this incident represents, arguing about that seems irrelevant."
Gene Allen, a fire protection specialist with Allianz Insurance, said that open air ignition of flammable vapors have not resulted in detonations.
"There are certain incidents where the vapors got into structures and other confined spaces, then when the flame front got to them it caused a point detonation," Allen said.
Months of intense investigation have isolated several key factors that contributed to the disaster at Buncefield. Both White and Allen have been analyzing the evidence to determine what happened.
Two obvious points of detonation were a fire pump house and a 40-foot cargo container that had been altered to house a generator. Both show signs of having a large internal overpressure, or explosion inside.
"The word 'detonation' is associated with TNT and explosives where the temperature generated gets to 5,000 or 6,000 degrees F, creating a very sharp explosive overpressure," Allen said. "Hydrocarbons, when they create an overpressure, the blast wave is much smaller in size and psi. The overpressure has a longer duration."
As compared to the demolished pump house and the generator container, the damage to many of the surrounding office buildings was due to the immense low pressure wave from the blast pushing through from one side to the other at the generator and 360 degrees from the pump house. The metal generator structure, though damaged, did remain intact.
White believes the intensity of the overpressure across the open area was largely the result of heavy humidity, lack of wind and low temperature.
"We had a detonation of this vapor cloud," White said. "Some people are saying the overpressure was as high as 15 pounds per square inch."
The quantity of spill to gasoline generated at Buncefield was hundreds of times larger than normally generated by a typical gasoline overfill, White said. As a result, damage from the overpressure was enormous compared to what would normally be seen.
"We had a couple of buildings or structures that had vapor in them when they blew and it blew the structures away," White said. "Most importantly, we had an open area where the vapor cloud created a overpressure that badly damaged a number of buildings within a 1,500 foot radius."
Since no means exist to control the degree of overpressure in such circumstances, the more effective area of research is making sure such overfills never occur again. At Buncefield, the fail safe systems apparently needed better fail safes themselves, Allen said.
"They had a gauging problem," Allen said. "The liquid level gauge stuck at about 3 a.m. and the ignition occurred about 6 a.m. with seemingly no human intervention. The secondary safety system which was the high level alarm and shutdown function appears to have been in the bypass position or just not working."
The bypass allowed operators to test event switches and other devices without fear of tripping a false alarm. However, it appeared there was no device to warn operators that the safety system may not have been in the normal position.
"Ideally, the disable device on the safety system would have required a key or a computer password so the operator knows that no one can simply turn it off when they want to," Allen said. "That person would have to get permission from the supervisor and that supervisor would have to use his key or password to do it."
As for other warning devices, White said he recommends that every storage tank receiving product from a pipeline have a high level alarm, an extremely high level alarm and an automatic shutoff.
"Not everyone has this," White said. "You can have the devices and technology, but if it is not maintained and properly operated you can have one of these disasters."
At Buncefield, the gasoline storage tank in question overflowed for around 60 minutes before the vapor cloud ignited.
In July, the U.K.'s Health and Safety Executive (HSE) issued a safety alert to operators of fuel storage facilities to carry out immediate checks where TAV level switches manufactured and supplied by Cynergy 3 Components Limited or their predecessor companies are fitted to their storage tanks.
"Operators should check TAV level switches fitted with change over (double throw) reed switches," the safety alert states. "These switches are normally used in 'de-energize' mode to maintain alarm/trip circuits in 'normal' status."
The switches are tested by using a lever or plate fitted to the head of the switch, which can be raised to simulate a high level of liquid in the tank, the report states. If the switch is working, then alarms and trips connected to the switch should operate.
"However, it is critical that after carrying out this test that the lever or plate is returned to the correct position and locked into place, using a special padlock supplied by the manufacturer and in accordance with the manufacturers instructions," the report states. "Failure to do this can lead to the switch being inoperative in normal operating mode even though it gives the appearance of functioning normally when tested."
Following the explosion was the emergency response. One key decision that needed to be made from the outset was continually deferred - Do we need to put this fire out?
"The normal fire department responses is 'We're going to fight it!' Allen said. "What they should have done is cover the exposures and then go back to the coffee shop to decide whether to put it out."
White said that maintaining control over the diverse elements of an industrial emergency response is difficult under the best conditions.
"It's not as if you can stop and say 'Okay, guys, what kind of foam do we have? What kind of delivery devices? What is the application rate we need?' Then when we figure that out we position the devices and supply them with water and foam.
"But that's not what happens at these things," White said. "You have an industrial truck on the eastside putting protein foam on it, you have a municipal truck on the Westside using High Ex foam, then you have another truck on the southside and he's using something else. Then, as you're fighting fire, you notice the foam coming out of your nozzle goes from brown to white to brown. You can't put out these big fires by just hollering 'Y'all come.' Nobody normally has a grasp on what is going on."
Having committed to putting the fire out, the firefighters at Buncefield then discovered that they would not be the ones in charge of the fire scene.
"For 16 hours the environmental people said 'You can't put a drop of water on it until you come up with a plan or method of capturing every drop of foam or water you put on it," White said. "That gave the fire 16 hours to decide what it was going to burn up."
Actually, the fire never spread much beyond the area claimed in the first three hours of the emergency. Along with the tank fires, firefighters were greeted with a host of pressure fires and dike fires as well.
"After a valve burns for 12 to 24 hours the valve seals leak," Allen said. "There is no way to shut it off. You could put out the tanks, but the product left in the tanks would still feed the pressure fires. Once valve damage occurs the only way to deal with it is pump foam subsurface in the line that is leaking and try to wrap something around the valve."
Three days later, the fires were still burning. The waste water that the environmentalists were so worried about still reached the nearby ground water and surface water, White said.
Nearly 200,000 gallons of foam were applied during the fire. Allen said firefighters failed to use aspirating nozzles when applying protein-based foams to the fire.
"They could probably have put it out even with foams like that if they could have gotten the application rate high enough," Allen said.
Switching to high expansion foam was equally unsuccessful.
"It just blew over the dikes and kept on going," Allen said.
As with fire water runoff, environmental politics factors into the decisions made about purchasing fire fighting foam in the U.K., White said.
"There is a big deal being made about fluorinated surfactants and PFOS in foam," White said. "The bottom line is we are creating a political atmosphere in Europe. The foams they were using and plan to use in the future - most of it barely works. Thank God in America we can still use real foam on fires."