HSE SRO :: The Buncefield Explosion Mechanism


The Buncefield Explosion Mechanism
	The Buncefield explosion (11 December 2005) resulted in tremendous damage to the outlying area and huge fires involving 23 large oil fuel tanks. One important aspect of the incident was the severity of the explosion, which would not have been anticipated in any major hazard assessment of the oil storage depot before the incident.



	On the recommendation of the Buncefield Major Incident Investigation Board (MIIB) Advisory Group, a Joint Industry Project (JIP) was undertaken to provide an understanding of the explosion mechanism and to define the scope of further work that may be necessary based on the findings. The Buncefield Explosion Mechanism Project (Phase 1) was jointly funded by the Health and Safety Executive (HSE), the UK Petroleum Industry Association (UKPIA), the Ministry of Housing of the Environment and Spatial Planning (The Netherlands), StatoilHydro and the Energy Institute.



	The findings from this project were disseminated for the first time at the Fire and Blast Information Group (FABIG) Technical Meetings on the 23rd and 24th of June 2009 in London and Aberdeen, and via a live interactive webcast. These meetings attracted the largest audience since the creation of FABIG with a record 350 delegates over the two days.



	Five speakers from the project Technical Group (Vincent Tam (BP), Jonathan Puttock (Shell Global Solutions), Mike Johnson (GL Industrial Services), Laurence Cusco (HSL), Ian Barnes (MoD) and Bassam Burgan (SCI)) gave presentations describing the forensic evidence, deflagration and detonation hypothesis, alternative mechanisms, assessment of structural damage and proposals for further work. These presentations were based on the project report which was recently published by HSE (RR718) and is available for download from the HSE and FABIG websites (http://www.hse.gov.uk/research/rrhtm/rr718.htm).



	Unlike other well known incidents, a vast amount of data in the form of witness statements, photographs, CCTV and video footage was available from the Buncefield incident, which enabled a detailed study of the explosion characteristics. It is estimated that a pancake shaped vapour cloud covering an area of around 120,000m2 with an average height of around 2m formed, and evidence suggest that the ignition source was in the emergency pump house (see Figure). High levels of overpressure (>200kPa) were seen in all areas within the cloud irrespective of the nature of the terrain and amount of congestion; the maximum overpressure was probably much higher.



	However, the overpressure diminished rapidly with distance away from the edge of the cloud and evidence suggests overpressures in the region of 5-10kPa within 150m. Another distinctive feature was the direction of net impulse (directional damage) which, within the cloud, acted in the direction opposite to the direction of the explosion propagation, whereas outside the cloud, acted in the direction of the explosion propagation.



	The deflagration and detonation explosion mechanisms were assessed for their consistency with the observed explosion characteristics. Deflagration was found to be inconsistent with the near-field significant damage to objects and cars, and with the net impulse as shown by directional damage to lamp posts, trees and posts within the flammable cloud.



	Detailed modelling of the area surrounding the emergency pump house supports the proposition that the trees and undergrowth along Three Cherry Trees Lane (see Figure) caused flame acceleration to a velocity of several hundred m/s, speed at which a transition to detonation is considered possible. Detonating vapour clouds are known to generate high overpressures and modelling confirmed the direction of the net impulse both within and outside the cloud and the rapid decline in overpressure with distance from the edge of the cloud.



	However, the magnitude of the loading predicted by the model would have caused greater damage to the Northgate and Fuji buildings. This discrepancy needs further investigation. This might be explained if, for example, the detonation was confined to part of the cloud depth (e.g. if only part of the cloud depth was at a concentration within detonability limits or if the cloud in some parts was not deep enough to sustain a detonation).



	From the work undertaken, the most likely scenario can be summarised as follows: • Dense vapour dispersion in very low wind speed conditions leading to a cloud build-up over an area of 120,000m2; • Ignition at the emergency pump house; failure of the pump house structure followed by a deflagration outside the pump house and flame propagation to the undergrowth and trees; • Flame acceleration in the undergrowth and trees along Three Cherry Trees Lane up to flame velocities of several hundred m/s, followed by a transition to detonation near the junction between Three Cherry Trees Lane and Buncefield Lane; • Detonation of part of the remaining gas cloud.



	This phase of the project has also defined experimental and analytical work to be completed in a second joint industry funded phase of the project. The aim of the proposed work (which is expected to take 24 months to complete) is to improve the understanding of low wind speed dispersion and large shallow vapour cloud explosions and the response of various forms of construction to such explosions. This phase has also identified areas for fundamental research of a longer term nature suitable for academic institutions and specialist technical organisations. The latter will be progressed through the normal funding channels for academic research (e.g. EPSRC, European Framework, etc.).



	The presentations generated a very high interest from the audience, which was particularly looking for practical advice following the work undertaken. A key point is to keep an open mind in a risk assessment when identifying potential sources of congestion as it is believed that the trees and bushes along Three Cherry Trees Lane played a critical role in accelerating the flame. It was also suggested that the focus should be on preventing losses of containment.



Contact Details :
Dr Bassam Burgan Deputy Director SCI Silwood Park Ascot Berkshire SL5 7QN Tel: +44 (0) 1344 636 545 Fax: +44 (0) 1344 636 570 Email: b.burgan@steel-sci.com

See also :
RR718 - Buncefield Explosion Mechanism Phase 1 (Volumes 1 and 2) research report


User Comments
User	Comment	Date Added

To be notified when more articles like this are released, please register for free.

Return