Buncefield Oil Depot Research Articles

Time reversal for localization of sources of infrasound signals in a windy stratified atmosphere.

Time reversal is used for localizing sources of recorded infrasound signals propagating in a windy, stratified atmosphere. Due to the convective effect of the background flow, the back-azimuths of the recorded signals can be substantially different from the source back-azimuth, posing a significant difficulty in source localization. The back-propagated signals are characterized by negative group velocities from which the source back-azimuth and source-to-receiver (STR) distance can be estimated using the apparent back-azimuths and trace velocities of the signals. The method is applied to several distinct infrasound arrivals recorded by two arrays in the Netherlands. The infrasound signals were generated by the Buncefield oil depot explosion in the U.K. in December 2005. Analyses show that the method can be used to substantially enhance estimates of the source back-azimuth and the STR distance. In one of the arrays, for instance, the deviations between the measured back-azimuths of the signals and the known source back-azimuth are quite large (-1° to -7°), whereas the deviations between the predicted and known source back-azimuths are small with an absolute mean value of <1°. Furthermore, the predicted STR distance is off only by <5% of the known STR distance.

A framework for minimizing domino effect through optimum spacing of storage tanks to serve in land use planning risk assessments

The recent interest in optimizing land use through integrated risk assessment, which intensified after Buncefield oil depot incident in 2005, calls for, among other things, determination of domino effects, which can be severe. The recommended tanks spacing and water application rates by design codes vary widely and are sometimes contradicting, if not subjective. This paper introduces a novel framework to determine the water application rate for protection of storage tanks against thermal radiation from an external non-contacting fire through first principles modeling. This new approach has been applied to assess the appropriate cooling water rate needed for the protection of an existing crude oil tank farm which includes three one-million barrel and two 500,000-barrel floating roof tanks. The tanks are to be protected from the thermal radiation of an adjacent tank with a full surface fire by application of cooling water and in the present arrangement they are so widely spaced that this is only attributable to a generous ‘overdesign’. It has been shown that applying the new approach could have resulted in at least 25% saving in tank farm area.

A Model for Temperature Fluctuations in a Buoyant Plume

We present a hybrid Lagrangian stochastic model for buoyant plume rise from an isolated source that includes the effects of temperature fluctuations. The model is based on that of Webster and Thomson (Atmos Environ 36:5031–5042, 2002) in that it is a coupling of a classical plume model in a crossflow with stochastic differential equations for the vertical velocity and temperature (which are themselves coupled). The novelty lies in the addition of the latter stochastic differential equation. Parametrizations of the plume turbulence are presented that are used as inputs to the model. The root-mean-square temperature is assumed to be proportional to the difference between the centreline temperature of the plume and the ambient temperature. The constant of proportionality is tuned by comparison with equivalent statistics from large-eddy simulations (LES) of buoyant plumes in a uniform crossflow and linear stratification. We compare plume trajectories for a wide range of crossflow velocities and find that the model generally compares well with the equivalent LES results particularly when added mass is included in the model. The exception occurs when the crossflow velocity component becomes very small. Comparison of the scalar concentration, both in terms of the height of the maximum concentration and its vertical spread, shows similar behaviour. The model is extended to allow for realistic profiles of ambient wind and temperature and the results are compared with LES of the plume that emanated from the explosion and fire at the Buncefield oil depot in 2005.

Dypasi Methodology: from Information Retrieval to Integration of Hazid Process

The availability of hazard identification methodologies based on early warnings is a crucial factor in the prevention of major accidents. Accidents like Seveso, Buncefield and Toulouse, where severe consequences occurred seemingly unexpected by the plant safety management, have made it clear that a comprehensive and complete identification and assessment of potential hazards in the process industry are of primary importance for the prevention and the mitigation of accident scenarios. The accident scenarios deviating from normal expectations of unwanted events or worst-case reference scenarios captured by common HAZard IDentification (HAZID) techniques are usually defined as “atypical”. The main issue posed by the prevention of atypical scenarios is the availability of techniques able to identify them within a routine HAZID process, capturing evidence of new hazards and learning from ‘early warnings’ as soon as they come to light. For this reason a specific method named Dynamic Procedure for Atypical Scenarios Identification (DyPASI) was developed. The method was conceived as a development of conventional bow-tie identification techniques. DyPASI is a method for the continuous systematization of information from early signals of risk related to past events. It dynamically integrates in the bow-ties the results of information retrieval activities. DyPASI features as a tool to support emerging risk management process, having the potentiality to contribute to an integrated approach aimed at breaking “vicious circles”, helping to trigger a gradual process of identification and assimilation of previously unrecognized atypical scenarios. The current contribution presents the technique and demonstrates the application by a selected case-study of practical application (Toulouse AZF accident, Buncefield oil depot accident, LNG regasification terminal, CCS plant).

Using plume rise schemes to model highly buoyant plumes from large fires

The atmospheric dispersion model Numerical Atmospheric-dispersion Modelling Environment (NAME) is used to simulate the smoke plume from the explosion at the Buncefield oil depot. Simple modelling, in which the plume rise is included through an effective elevated source term, captures the transport and spread of the plume well. More complex modelling, using the NAME plume rise scheme, underestimates the plume rise and plume vertical spread. We consider a number of potential reasons for this underprediction and compare NAME predictions against large-eddy simulations (LES) of the plume.

Forensic investigation of blast damage to British buildings

Buildings in mainland Britain have suffered from several serious blast incidents over the past two decades, from the 1992 St Mary Axe bomb in the City of London to the Buncefield oil depot explosion in 2005. This paper is based on the authors' experiences of investigating and repairing the resulting damage to many of the major buildings involved in these incidents. It describes how forensic investigation – including laboratory and on-site testing and computer modelling and analysis techniques – was vital to these investigations to assess dynamic response, extent of damage and residual strength of the buildings and structural elements.

The Buncefield explosion: a benchmark for infrasound analysis across Central Europe

SUMMARY A large explosion occurred at the Buncefield Oil Depot, United Kingdom, on the 2005 December 11, as the result of a spillage of 300 tonnes of petrol. The explosion generated large amplitude infrasonic signals that were recorded at ranges of up to 1400 km. We present an analysis of data from across Central Europe using seven infrasound arrays, four of which recorded a total of 20 infrasonic arrivals, and 49 seismometers, which recorded 89 air-toground coupled arrivals. Exceptionally high amplitude stratospheric winds occurred at the time of the explosion, propagating east–southeast with speeds of up to 130 m s −1 , generating a highly efficient duct for infrasound propagation. The Buncefield explosion can therefore be considered as a benchmark for studying the influence of different propagation algorithms and atmospheric models on phase identification, source location and yield estimation. Significant improvements are made in phase identification using accurate atmospheric parametrizations and 3-D modelling, where we identify 18 of the 20 observed phases, compared to seasonal atmospheric models and 1-D ray tracing which only identify seven of the phases. Additionally, various location scenarios are considered, including cross-bearings and least-squares traveltime and backazimuth inversion, all of which estimate locations within 35 km of ground truth. A novel single array location method using multiple phases works comparably well. Yield estimates made using the stratospheric arrival amplitudes exhibit one order of magnitude variability across the network, with a mean estimate of 51 tonnes high explosive equivalent. Detailed studies, such as this, are important for developing a successful monitoring regime for atmospheric or surface events, and for assessing the capability of the infrasound monitoring component of the Comprehensive Nuclear-Test-Ban Treaty verification system.

Large-eddy simulation of the plume generated by the fire at the Buncefield oil depot in December 2005

The explosion at the Buncefield oil depot in Hertfordshire, UK on 11 December 2005 produced the largest fire in Europe since the Second World War. The magnitude of the fire and the scale of the resulting plume thus present a stringent test of any mathematical model of buoyant plumes. A large-eddy simulation of the Boussinesq equations with suitable initial conditions is shown to reproduce the characteristics of the observed plume; both the height of the plume above the source and the direction of the downwind spread agree with the observations. This supports the use of the Boussinesq assumption, even for such a powerful plume as the one generated by the Buncefield fire. The presence of a realistic water vapour profile does not lead to significant additional latent heating of the plume, but does lead to a small increase in the final rise height of the plume due to the increased buoyancy provided by the water vapour. Our simulations include the interaction of radiation with the aerosol in the plume, and reproduce the observed optical thickness of the plume and the reduction of solar radiation reaching the ground. Far downwind of the source, solar radiation is shown to play a role in lofting the laterally spreading plume, but the manner in which it does so depends on the aerosol concentration. In the case of high aerosol concentration, the thickness of the plume increases; the incoming solar radiation is absorbed over such a small depth that only the top of the plume is lofted upwards and the level of maximum concentration remains almost unchanged relative to the case with no radiation. When the aerosol concentration is low, the whole plume is heated by the incoming solar radiation and the lofting is more coherent, with the result that the level of maximum concentration increases relative to the case with no radiation, but the thickness of the plume increases only slightly.

The influence of shallow atmospheric structure on tropospheric ducted infrasound from the Buncefield oil depot explosion

The vapour cloud explosion which destroyed the Buncefield oil depot, UK, on 11th December 2005, has proven to be a benchmark ground truth event for infrasonic studies. The regional infrasonic returns, those that propagated in the stratosphere and thermosphere, have been analysed in detail elsewhere. Here, we present the results of a study into infrasound ducted in the troposphere, recorded within 250 km of the source as air-to-ground coupled waves by a dense seismometer network. These tropospheric arrivals exhibit large waveform differences across the UK, both in amplitude and waveform shape. We numerically model these infrasound arrivals using a wave number integration method, incorporating a velocity profile derived from the UK Met Office numerical weather prediction model. Although some of the waveform variability is due to ground conditions at the recording site, we show that consistent changes in waveform shape across a 200 km swath of stations are correlated with a change in the wind vector in the lowermost 2 km of the atmosphere. Also, small amplitude, high-frequency precursors to the dominant acoustic signal, which might be misinterpreted as evidence of a small initial explosion, are shown to be consistent with the dispersion expected from a thin, shallow wind jet.

The Buncefield oil depot explosion: where there's smoke, there's (stock market) fire?

This study examines the stock market response to the Buncefield oil depot explosion in 2005. Like previous studies on technological disasters, we find an adverse effect on security prices. However, average abnormal return is only −0.58% for the four oil firms involved in the accident; that is, the explosion did not throw shareholders into panic selling.

Seismo-acoustic analysis of the Buncefield oil depot explosion in the UK, 2005 December 11

SUMMARY A massive vapour cloud explosion occurred at the Buncefield fuel depot near Hemel Hempstead, UK, in the morning of 2005 December 11. The explosion was the result of an overflow from one of the storage tanks with the release of over 300 tons of petrol and generating a vapour cloud that spread over an area of 80 000 m2, before being ignited. Considerable damage was caused in the vicinity of the explosion and a total of 43 people were injured. The explosion was detected by seismograph stations in the UK and the Netherlands and by infrasound arrays in the Netherlands. We analysed the seismic recordings to determine the origin time of 06:01:31.45 ±0.5 s (UTC) from P-wave arrival times. Uncertainties in determination of origin time from acoustic arrival times alone were less than 10 s. Amplitudes of P-, Lg and primary acoustic waves were measured to derive decay relationships as function of distance. From the seismic amplitudes we estimated a yield of 2–10 tons equivalent to a buried explosion. Most seismic stations recorded primary and secondary acoustic waves. We used atmospheric ray tracing to identify the various travel paths, which depend on temperature and wind speed as function of altitude, leading to directional variation. Refracted waves were observed from the troposphere, stratosphere and thermosphere with a good match between observed and calculated traveltimes. The various wave types were also identified through array processing, which provides backazimuth and slowness, of recordings from an infrasound array in the Netherlands. The amplitude of stratospheric refracted acoustic waves recorded by the array microbarometers was used to estimate a yield of 21.6 (±5) tons TNT equivalent. We have demonstrated through joint seismo-acoustic analysis of the explosion that both the seismic velocity model and the atmospheric model are sufficient to explain the observed traveltimes.

英国バンスフィールド油槽所で発生した爆発火災について

2005年12月11日に英国ヘメル・ヘムステッドのバンスフィールド油槽所で発生した爆発火災について，英国HSEと英国環境庁が実施した事故調査の第一報を日本語訳して要約したものを紹介する．本災害のような非常に大規模な爆発火災は，幸い日本では近年起きていないが，その現象について理解を深めることは，規模は違っても類似災害をふせぐ一助になると考える．

The Buncefield oil depot incident: a discussion of the meteorology

The Buncefield oil depot incident: a discussion of the meteorology

The dispersion of the Buncefield oil fire plume: An extreme accident without air quality consequences

The dispersion of pollutants from the huge Buncefield oil depot fire that occurred on 11 December 2005 is simulated using a regional Eulerian chemistry-transport model. We analyse the transport and mixing of the fire plume. We show that the hot plume never reached the ground. Instead, it pierced the thin wintertime boundary layer and was injected into the free troposphere at higher altitudes. This is in agreement with data from many air quality stations. This high injection was fortunate because the fine aerosol particles (PM 10) mass column generated by the fire smoke exceeded that of ordinary pollution by an order of magnitude. Our regional chemistry-transport modelling is able to predict the early development of the plume dispersion, as shown by a qualitative comparison between simulated PM 10 columns and a satellite image obtained by the EOS-TERRA-MODIS sensor. If the accident had occurred in summer when boundary layers are much deeper and convective, a severe degradation in air quality due to PM 10 could have occurred, as shown by a sensitivity simulation assuming a similar fire during one of the hottest days of August 2003. The modelled impact of the fire on regional and European air quality levels strongly depends on the altitude reached by the buoyant plume, as shown by a set of sensitivity simulations with variable injection heights. However, in all cases we found that the fire only affected surface aerosol concentrations without increasing photochemical pollution.

Health impact of the Buncefield oil depot fire, December 2005: Study of accident and emergency case records

On Sunday 11th of December 2005, an explosion occurred at the Buncefield oil depot, Hertfordshire, resulting in a large fire that blazed for several days. Two Accident and Emergency (A&E) departments were placed on stand-by to receive casualties. A simple surveillance system was put in place during the acute phase of the incident, but this was not adequate to fully examine the health impact of the fire. Retrospective study of A&E records at Hemel Hempstead and Watford A&E departments for the period that the fire burnt. Two hundred and forty-four people attended A&E as a result of the fire: 187 were members of the emergency services; 17 were oil depot workers; and 40 were members of the public. The most common presenting complaints were respiratory symptoms (n = 66) and injuries (n = 38). Twenty-five (21%) individuals were referred for medical follow-up. There were no fatalities. There was a significant impact on local health services, with many emergency service personnel attending A&E asymptomatically. Alternative health service provision for these individuals, possibly near to the scene of the incident, should be considered. This incident has also highlighted an urgent need to develop surveillance systems that enable real-time monitoring of the acute public health impact of major incidents.

Observations of the plume generated by the December 2005 oil depot explosions and prolonged fire at Buncefield (Hertfordshire, UK) and associated atmospheric changes

The explosions and subsequent fire at the Buncefield oil depot in December 2005 afforded a rare opportunity to study the atmospheric consequences of a major oil fire at close range, using ground-based remote-sensing instruments. Near-source measurements (less than 10 km) suggest that plume particles were approximately 50% black carbon (BC) with refractive index 1.73−0.42i, effective radius ( R eff ) 0.45–0.85 μm and mass loading approximately 2000 μg m −3 . About 50 km downwind, particles were approximately 60–75% BC with refractive index between 1.80−0.52i and 1.89−0.69i, R eff ∼1.0 μm and mass loadings 320–780 μg m −3 . Number distributions were almost all monomodal with peak at r &lt;0.1 μm. Near-source UV spectroscopy revealed elevated trace gas concentrations of SO 2 (70 ppbv), NO 2 (140 ppbv), HONO (20 ppbv), HCHO (160 ppbv) and CS 2 (40 ppbv). Our measurements are consistent with others of the Buncefield plume, and with studies of the 1991 Kuwaiti oil-fire plumes; differences from the latter reflecting in part contrasts in combustion efficiency and source composition (refined fuels versus crude oils) leading to important potential differences in atmospheric impacts. Other measurements made as the plume passed overhead approximately 50 km downwind showed a reduced solar flux reaching the surface, but little effect on the atmospheric potential gradient (electric field). The wind speed data from the day of the explosion hint at a possible explosion signature.

A national syndromic surveillance system for England and Wales using calls to a telephone helpline

Routine primary care data provide the means to monitor a variety of syndromes which could give early warning of health protection issues. In the United Kingdom, a national syndromic surveillance system, operated jointly by the UK Health Protection Agency (HPA) and NHS Direct (a national telephone health helpline), examines symptoms reported to NHS Direct. The aim of the system is to identify an increase in syndromes indicative of common infections and diseases, or the early stages of illness caused by the deliberate release of a biological or chemical agent. Data relating to 11 key symptoms/syndromes are received electronically from all 22 NHS Direct call centres covering England and Wales and analysed by the HPA on a daily basis. Statistically significant excesses in calls are automatically highlighted and assessed by a multi-disciplinary team. Although the surveillance system has characterised many sudden rises in syndromes reported to NHS Direct, no evidence of a biological or chemical attack has been detected. Benefits of this work, however, are early warning and tracking of rises in community morbidity (e.g. influenza-like illness, heatstroke); providing reassurance during times of perceived high risk (e.g. after the 7 July 2005 London bombs and December 2005 Buncefield oil depot fire); and timely surveillance data for influenza pandemic planning and epidemic modeling.