The UK Buncefield Incident - The View From the ‘Middle,’ That of a UK Risk Assessment Engineer

摘要

The large scale release and subsequent massive explosion of petrol from a storage tank at the Buncefield storage site in December 2005 has resulted in two court cases (Civil and Criminal) and in significant ongoing learning for the UK risk assessment community. This paper presents the personal views and lessons learnt by a UK based risk assessment consultant engineer, both from involvement as an expert witness in the civil case and in support of storage tank operators post incident. The view is from an engineer typically being asked to support operations using accepted typical assessment methods and practices to develop the likes of safety cases and design reviews. This is the 'middle' ground, as these views fit in between those developed by academic and technical (experimental) research organisations. Subjects covered are - The lack of poor UK awareness of previous large scale petrol-based explosions (does this apply elsewhere in the world?); additional factors for consideration in risk assessment by improved hazard identification and determination of the scale of potential consequences, such as zero wind conditions and presence of minor factors in explosion generation that would previously have been discounted; the operation of typical tank storage; and, improvements to the design of prevention and mitigation measures for tank storage. Some of the key learning from the Buncefield incident include: 1. Trips and alarms should be SIL reviewed and maintained as appropriate (correctly according to OEM details) 2. The lack awareness and standard assessment of previous incidents - another case of incidents not being properly reviewed and retained in corporate or even in cultural memory (effort is typically put in to investigate accident sequence; little effort is undertaken to confirm that say consequence size would have been correctly determined from available methodologies) 3. There should be a form of accountancy system in place on tank filling - e.g., expectation of level changes on volume transfers - tank filling should not be controlled solely on level instrumentation (for example pump rates and running duration should be matched to level changes). 4. Tank bund design should be improved - No pipe penetrations, sumps should be pumped out, seals should be fire resistant, allowance for foam application. 5. Hydrocarbon leak and vapour detection is important (now critical if ignition sources are being controlled via ATEX/DSEAR reviews) as small/medium releases can collect to form large clouds by the time of ignition. The larger consequence significantly outweighs any likelihood reduction from control over ignition source, if it remains undetected. (There was a 40 minute delay before ignition at Buncefield). Greater focus on HC leak detection and isolation ability. 6. The placing of emergency equipment should be reviewed - at Buncefield the fire water pumps, although outside the hazardous area were destroyed in the blast. 7. Risk assessment - The potential for very still weather conditions should be reviewed for dense/neutrally buoyant gas/vapour dispersion. 8. Risk assessment - The potential for significant blast overpressure should be considered for moderately reactive materials - e.g., Butane, Propane and Pentane and factors such as vegetation should be seen as a mechanism for turbulence (this should also cover temporary turbulence factors such as scaffolding as the likelihood for this to be present is of the same order as say low frequency zero wind conditions). The Potential maximum flame length should also be considered, so that long thin congested regions are picked up as possible significant overpressure generation mechanism. 9. Risk assessment - The potential for Bang Box jet ignition of a large vapour cloud should be reviewed. This is where initial ignition may occur in small enclosed spaces and the resultant flame jets into the unconfined vapour cloud. The increased ignition energy from the ‘bang box’ could for moderatel