Is All the Post-Buncefield Research a Waste of Money?

It is now several years since I spent most of one December Sunday in 2005 watching the television coverage of the Buncefield fire on the BBC News Channel (or BBC News 24 as it was then called). Although I now live in Wirral, NW England, I lived for 5 years in a village on the outskirts of Hemel Hempstead, although on the other side of the town centre from Buncefield. The Buncefield depots were a familiar sight as I drove to the M1 motorway on my way to work in Hitchin and later Luton. Now that we are approaching the fifth anniversary of the event it is perhaps time to reflect on the various regulatory activities, and other activities promoted by the regulators.

Unsurprisingly, the first reaction of the joint competent authority was to start an investigation into the cause(s) of the incident. As the incident had received such overwhelming media coverage an independent investigation board (The Buncefield MIIB) was set up, chaired by Lord Tony Newton, a former minister in a previous Conservative government. The investigation board soon widened their focus to include the issue of land use planning (LUP) advice given around the Buncefield sites over the previous years, and the policies and practices used by HSE in order to produce that advice. Less well know is the fact that HSE also initiated an internal 'policy and procedures' study which essentially reported on what HSE's stated approaches were, to regulation at Buncefield, and whether these approaches had actually been followed in practice. The current author was a member of that study team. As far as I can tell that report remains unpublished.

During the course of the investigation it rapidly became clear that much of the fire and explosion modelling used by HSE for LUP, and duty holders for regulatory compliance, had not predicted what had actually happened in the incident. I do not regard this as surprising because loss of containment of flammable substances can lead to a range of possible outcomes according to a wide range of ever changing conditions, that might apply on the day of the incident. To deal with the complex issues of explosion modelling, the Buncefield MIIB brought together an Explosion Mechanism Advisory Group which duly reported in August 2007. The conclusions in this 32 page report are reproduced below:

Conclusions

13 The Group attempted to explain the explosion event at Buncefield using deflagration, detonation or a combination of both. It also examined other possible means of flame acceleration. However, it was not possible to identify a single scenario that could explain all aspects. In this the Group was limited by:

uncertainty regarding the composition of the vapour cloud;
apparent ambiguities in some of the forensic evidence;
uncertainty regarding the explosion severity required to cause the level of damage observed, particularly to cars and buildings;
the time available, as the possible scenarios could not be properly tested against the considerable amount of information available; and
the difficulty of distinguishing between unburned gas flow ahead of the flame and burned gas flow, in the other direction, away from it.

14 Nevertheless, the Group concluded that there is a strong likelihood that the cause of the severe explosion at Buncefield can be explained. However, this will require further, more detailed work. It is the opinion of the Group that this should in the first instance involve the continuation of the assessment already started, fully testing a range of scenarios against the forensic evidence and the current scientific understanding of explosion mechanisms.

15 It is also the opinion of the Group that a full explanation is unlikely to be achieved without the conduct of further research.

This highlights the problems of explosion modelling even where most of the conditions at the time of the explosion are known from the detailed results of the investigation. Put yourself in the position of HSE staff and duty holders who face the challenge of explosion modelling prior to the event. Which of the many time varying conditions should they attempt to take into account? Does explosion modelling knowledge allow them to be taken into account even if the attempt is to be made?

In a chaotic world (in the scientific sense) can anyone be sufficiently sure about the scope and accuracy of our knowledge of the starting conditions to reliably predict the outcome? It seems to me that a good analogy can be drawn with the world's meteorological organisations and their modelling of the weather. I will not try to compare and contrast the accuracy of weather and gas explosion predictions, but it seems to me that when one considers the importance of weather forecasting and the funds expended on it, reaching a similar standard with gas explosions will require exceedingly large sums of money.

The Buncefield Explosion Mechanism Advisory Group made a recommendation that further research work should be carried out to 'determine the explosion mechanism in the Buncefield incident'. The proposed 2 phases of work, of which the first was costed at £200,000. As phase two involves large scale experimental work to fill in gaps in understanding, the advisory group estimate that the cost will be in excess of £1 million. The objective is only to permit understanding of what happened, although it clearly will give additional modelling knowledge for future application by the modelling community if it has a clear understanding of what needs to be modelled.

When one looks back at the modelling that was carried out by HSE in order to give LUP advice, HSE settled on a protection concept approach(see footnote 1) with hazard ranges predicted from a 100 metre diameter pool which had over-topped the bund. Such a pool fire would have been modelled as giving flames 200 metres high. If you place this pool to the west of tank 912 you will find that it covers virtually the whole of the Northgate building car park with the building itself almost in the burning pool. Picture the scene had this been the incident that had occurred. Would the outcome have been substantially different?

I have heard nothing further about a Joint Industry Project as envisaged by the Buncefield Explosion Mechanism Advisory Group. If it goes ahead, I wonder if it will be money well spent?

(1) See Advisory Committee on Major Hazards, Third Report, Paragraph 82 'Ideally, the separation should be such that the population would be unaffected whatever accident occurs. For hazardous installations, however, such a policy is not reasonably practicable. It seems reasonable to aim for a separation which gives almost complete protection for lesser and more probable accidents, and worthwhile protection for major but less probable accidents.'

This work is licensed under a Creative Commons Attribution-NoDerivs 2.0 UK: England & Wales License.