Buncefield blast concern
28 Nov 2006
Experts remain mystified by the ferocity of the explosion almost one year ago
The Buncefield incident remains of great concern to the safety investigators around the world, who remain mystified by the ferocity of the explosion that occurred at the fuel depot, near Hemel Hempstead on Sunday, 11 December last year.
None of the models, including those based on masses of WW2 ordinance data, nuclear explosions and more modern explosives, actually work for Buncefield, said Taf Powell, manager of the Health & Safety Executive's investigation into the incident.
"We are simply unable to explain this at the moment. The most violent explosion was in an almost totally open area, the HSE man continued. "Where you would have expected a fairly slow moving flame front with flame speeds up to about a metre/sec we had tremendously violent explosions with buildings significantly damaged and motor vehicles flattened."
Designers and planners have been taking the sensible worst-case scenario as the collapse of a large tank and a very large fire in the resulting pool of escaped fuel, and have based their planning advice on that scenario.
Buncefield investigators have since learned of other incidents around the world where ignition of an unconfined petrol vapour cloud has resulted in a very violent explosion. But, said Powell, there does not seem to have been any serious work to actually model this. The priority, said Powell, "is to develop adequate models to help designers and planners to get it right for the longer term, as the research will probably take some time."
Forensic experts at the HSE's research centre in Buxton are developing models of the fuel vapourisation process that created a virtually perfect flammable mixture by the time it got from the tank to the car parks at Buncefield. They are also trying to estimate the overpressure in the most violent part of the explosion — work that has involved putting cars inside explosion chambers to try to replicate the damage.
According to Powell, nothing could have prepared you for the scene of "absolute devastation" at the Buncefield site just after the explosion. When the investigators took control of the site on the following Thursday, conditions presented tremendous hazards to everyone, with a lot of dangerously stored fuel and many unstable structural elements. It was vital, he said, to ensure that anybody who needed to be around the site was working with as much protection as possible. The investigators then planned how to secure all the documentary evidence and physical evidence on the site.
Forensic experts began the difficult task of mapping and interpreting all the on-site and off-site damage, to begin to build a picture around the violent nature of the explosion. By 20 Dec, the investigators had all the computers in controlled storage to ensure the best possible chance of replicating and reading the data storage systems.
These are industrial data storage systems - very complicated systems that store weeks of colour CCTV footage, records of the fuel transactions into and out of the site, flow rates, valve positions temperatures and so on, said Powell.
"There was an awful lot of information to examine. And we had to find a way of rebuilding the discs and copying them in a way that made the data we received acceptable to a court if proceedings were to be taken in the future," he explained.
Perhaps the first big breakthrough for the investigators was in being able to view the visible mist spreading from bund A from CCTV data, which helped to confirm initial suspicions about the events that led up to explosion.
"There was some good CCTV footage looking straight down the bund wall. To see this mist flowing over the bund wall and quickly enveloping that part of the site and going off site was a Eureka moment," commented Powell.
In particular, the CCTV footage helped to focus the investigation in terms of the information needed from site operations, manufacturers and suppliers.
According to Powell, specialist engineers on the investigation did an outstanding job in replicating the control systems on the site, to enable knowledge of how the equipment could have worked in the first place, he said.
Another important development was the recovery of the level gauges and other equipment that failed to operate prior to fuel overflowing from the tank, leading to the formation of the exploding vapour cloud.
"We weren't able to near the tank until February as conditions were so dangerous," said Powell. With the help of a crane and inspectors — who didn't suffer from vertigo — the team was able to identify equipment inside the remains of tank 912 they had not previously expected to find. These items were sent off for forensic testing at Buxton — work that is still ongoing, often in presence of representatives of the equipment owners, the depot operators and others with an interest in the outcome of the investigation.
While Powell was restricted from commenting on many issues concerning Buncefield — legal and criminal proceedings are still under consideration — he did comment on other major industrial accidents worldwide.
Powell, who is head of HSE's Offshore division, joined the HSE from the oil industry in 1991 to develop the new offshore regulatory framework - a recommendation from the Piper Alpha disaster inquiry.
According to Powell, disasters such as Piper Alpha, as well as Flixborough, Ladbroke Grove and Texas City last year, show that the world continues to have accidents in situations where the risks of a major accident should by now be well controlled.
With most of these incidents a number of preventative systems seem to combine to bring about an accident, he said. "Professional engineers may say that none of the steps that contributed to most major incidents are in themselves remarkable, but are a combination of failures of protection systems that lined up to cause a serious accident."
Another feature of some previous incidents, said Powell, is "a culture that was strong in terms of occupational safety: slip trip and falls, protective equipment and so on but defective on the issue of controlling major hazard risks.
In the past some high hazard industries have accreted more layers onto their systems as regulatory requirements and operational standards change, but adding more barriers does not necessarily make you safer; it is the quality and robustness of those barriers." With very complicated site safety systems there can also be a loss of sight of the highest priority safety issues people need to be aware of, he added.
There are also common challenges in high hazard industries such as nuclear, petrochemical and oil & gas, according to the HSE expert. These, he said, include a trend towards managing significant parts of a business through contractors and having fewer of your own people directly on the front line.
Also ageing plants present challenges for maintaining safety critical equipment. "As a plant gets older it is less easy to maintain and it gets easier to infringe the original design intent when you do modifications. As well as this, companies may divest ageing assets to smaller companies that have not previously had to get used to running these facilities, some of which may not be functioning well". He continued: "These are all manageable, but they are nonetheless very persistent challenges."
None of the models, including those based on masses of WW2 ordinance data, nuclear explosions and more modern explosives, actually work for Buncefield, said Taf Powell, manager of the Health & Safety Executive's investigation into the incident.
"We are simply unable to explain this at the moment. The most violent explosion was in an almost totally open area, the HSE man continued. "Where you would have expected a fairly slow moving flame front with flame speeds up to about a metre/sec we had tremendously violent explosions with buildings significantly damaged and motor vehicles flattened."
Designers and planners have been taking the sensible worst-case scenario as the collapse of a large tank and a very large fire in the resulting pool of escaped fuel, and have based their planning advice on that scenario.
Buncefield investigators have since learned of other incidents around the world where ignition of an unconfined petrol vapour cloud has resulted in a very violent explosion. But, said Powell, there does not seem to have been any serious work to actually model this. The priority, said Powell, "is to develop adequate models to help designers and planners to get it right for the longer term, as the research will probably take some time."
Forensic experts at the HSE's research centre in Buxton are developing models of the fuel vapourisation process that created a virtually perfect flammable mixture by the time it got from the tank to the car parks at Buncefield. They are also trying to estimate the overpressure in the most violent part of the explosion — work that has involved putting cars inside explosion chambers to try to replicate the damage.
According to Powell, nothing could have prepared you for the scene of "absolute devastation" at the Buncefield site just after the explosion. When the investigators took control of the site on the following Thursday, conditions presented tremendous hazards to everyone, with a lot of dangerously stored fuel and many unstable structural elements. It was vital, he said, to ensure that anybody who needed to be around the site was working with as much protection as possible. The investigators then planned how to secure all the documentary evidence and physical evidence on the site.
Forensic experts began the difficult task of mapping and interpreting all the on-site and off-site damage, to begin to build a picture around the violent nature of the explosion. By 20 Dec, the investigators had all the computers in controlled storage to ensure the best possible chance of replicating and reading the data storage systems.
These are industrial data storage systems - very complicated systems that store weeks of colour CCTV footage, records of the fuel transactions into and out of the site, flow rates, valve positions temperatures and so on, said Powell.
"There was an awful lot of information to examine. And we had to find a way of rebuilding the discs and copying them in a way that made the data we received acceptable to a court if proceedings were to be taken in the future," he explained.
Perhaps the first big breakthrough for the investigators was in being able to view the visible mist spreading from bund A from CCTV data, which helped to confirm initial suspicions about the events that led up to explosion.
"There was some good CCTV footage looking straight down the bund wall. To see this mist flowing over the bund wall and quickly enveloping that part of the site and going off site was a Eureka moment," commented Powell.
In particular, the CCTV footage helped to focus the investigation in terms of the information needed from site operations, manufacturers and suppliers.
According to Powell, specialist engineers on the investigation did an outstanding job in replicating the control systems on the site, to enable knowledge of how the equipment could have worked in the first place, he said.
Another important development was the recovery of the level gauges and other equipment that failed to operate prior to fuel overflowing from the tank, leading to the formation of the exploding vapour cloud.
"We weren't able to near the tank until February as conditions were so dangerous," said Powell. With the help of a crane and inspectors — who didn't suffer from vertigo — the team was able to identify equipment inside the remains of tank 912 they had not previously expected to find. These items were sent off for forensic testing at Buxton — work that is still ongoing, often in presence of representatives of the equipment owners, the depot operators and others with an interest in the outcome of the investigation.
While Powell was restricted from commenting on many issues concerning Buncefield — legal and criminal proceedings are still under consideration — he did comment on other major industrial accidents worldwide.
Powell, who is head of HSE's Offshore division, joined the HSE from the oil industry in 1991 to develop the new offshore regulatory framework - a recommendation from the Piper Alpha disaster inquiry.
According to Powell, disasters such as Piper Alpha, as well as Flixborough, Ladbroke Grove and Texas City last year, show that the world continues to have accidents in situations where the risks of a major accident should by now be well controlled.
With most of these incidents a number of preventative systems seem to combine to bring about an accident, he said. "Professional engineers may say that none of the steps that contributed to most major incidents are in themselves remarkable, but are a combination of failures of protection systems that lined up to cause a serious accident."
Another feature of some previous incidents, said Powell, is "a culture that was strong in terms of occupational safety: slip trip and falls, protective equipment and so on but defective on the issue of controlling major hazard risks.
In the past some high hazard industries have accreted more layers onto their systems as regulatory requirements and operational standards change, but adding more barriers does not necessarily make you safer; it is the quality and robustness of those barriers." With very complicated site safety systems there can also be a loss of sight of the highest priority safety issues people need to be aware of, he added.
There are also common challenges in high hazard industries such as nuclear, petrochemical and oil & gas, according to the HSE expert. These, he said, include a trend towards managing significant parts of a business through contractors and having fewer of your own people directly on the front line.
Also ageing plants present challenges for maintaining safety critical equipment. "As a plant gets older it is less easy to maintain and it gets easier to infringe the original design intent when you do modifications. As well as this, companies may divest ageing assets to smaller companies that have not previously had to get used to running these facilities, some of which may not be functioning well". He continued: "These are all manageable, but they are nonetheless very persistent challenges."
