In-depth analysis of accidental oil spills from tankers in the context of global spill trends from all sources

This study gives a global overview of accidental oil spills from all sources (> or =700t) for the period 1970-2004, followed by a detailed examination of trends in accidental tanker spills. The present analysis of the number and volume of tanker spills includes temporal and spatial spill trends, aspects of spill size distribution as well as trends of key factors (i.e., flag state, hull type, tanker age, accident cause and sensitivity of location). Results show that the total number and volume of tanker spills have significantly decreased since the 1970s, which is in contrast to increases in maritime transport of oil and to popular perceptions following recent catastrophic events. However, many spills still occur in ecologically sensitive locations because the major maritime transport routes often cross the boundaries of the Large Marine Ecosystems, but the substantially lower total spill volume is an important contribution to potentially reduce overall ecosystem impacts. In summary, the improvements achieved in the past decades have been the result of a set of initiatives and regulations implemented by governments, international organizations and the shipping industry.     

Dispersion hazard from spills of LNG on land and on water

Previously published data on the dispersion of vapours from LNG spills on both land and water are correlated using common dispersion relationships     

Keys to modeling LNG spills on water

Although no LNG ship has experienced a loss of containment in over 40 years of shipping, it is important for risk management planning to understand the predicted consequences of a spill. A key parameter in assessing the impact of an LNG spill is the pool size. LNG spills onto water generally result in larger pools than land spills because they are unconfined. Modeling of LNG spills onto water is much more difficult than for land spills because the phenomena are more complex and the experimental basis is more limited.

The most prevalent practice in predicting pool sizes is to treat the release as instantaneous or constant-rate, and to calculate the pool size using an empirical evaporation or burn rate. The evaporation or burn rate is particularly difficult to estimate for LNG spills on water, because the available data are so limited, scattered, and difficult to extrapolate to the large releases of interest.

A more effective modeling of possible spills of LNG onto water calculates, rather than estimating, the evaporation or burn rate. The keys to this approach are to:

• Use rigorous multicomponent physical properties.

• Use a time-varying analysis of spill and evaporation.

• Use a material and energy balance approach.

• Estimate the heat transfer from water to LNG in a way that reflects the turbulence.

These keys are explained and demonstrated by predictions of a model that incorporates these features. The major challenges are describing the effects of the LNG–water turbulence and the heat transfer from the pool fire to the underlying LNG pool. The model includes a fundamentally based framework for these terms, and the current formulation is based on some of the largest tests to-date. The heat transfer coefficient between the water and LNG is obtained by applying a “turbulence factor” to the value from correlations for quiescent film and transition boiling. The turbulence factor is based on two of the largest unignited tests on water to-date. The heat transfer from the fire to the pool is based on the burning rate for the largest pool fire test on land to-date.     

A review of large-scale LNG spills: experiments and modeling

The prediction of the possible hazards associated with the storage and transportation of liquefied natural gas (LNG) by ship has motivated a substantial number of experimental and analytical studies. This paper reviews the experimental and analytical work performed to date on large-scale spills of LNG. Specifically, experiments on the dispersion of LNG, as well as experiments of LNG fires from spills on water and land are reviewed. Explosion, pool boiling, and rapid phase transition (RPT) explosion studies are described and discussed, as well as models used to predict dispersion and thermal hazard distances. Although there have been significant advances in understanding the behavior of LNG spills, technical knowledge gaps to improve hazard prediction are identified. Some of these gaps can be addressed with current modeling and testing capabilities. A discussion of the state of knowledge and recommendations to further improve the understanding of the behavior of LNG spills on water is provided.     

Potential for BLEVE associated with marine LNG vessel fires

Recent LNG marine shipping hazard studies have discounted BLEVE hazards associated with LNG vessels. This exclusion of a potential major hazard event has been queried, particularly since a recent LNG truck BLEVE-like event in Spain. This paper reviews the physical factors associated with the Spanish LNG truck event and accepts that this had features of a classical BLEVE event and that there is no inherent property of LNG excluding BLEVE-like events, although US LNG trucks would be safer due to design features. Marine LNG vessels have differently designed tanks and it is demonstrated that the combination of physical barriers makes direct thermal input to the LNG inner tank more limited than hypothesized by some, but if it occurs these tanks cannot rise to a pressure sufficient to cause a large flash of liquid and consequent BLEVE event of a scale hypothesized in the literature.     

An LNG release, transport, and fate model system for marine spills

LNGMAP, a fully integrated, geographic information based modular system, has been developed to predict the fate and transport of marine spills of LNG. The model is organized as a discrete set of linked algorithms that represent the processes (time dependent release rate, spreading, transport on the water surface, evaporation from the water surface, transport and dispersion in the atmosphere, and, if ignited, burning and associated radiated heat fields) affecting LNG once it is released into the environment. A particle-based approach is employed in which discrete masses of LNG released from the source are modeled as individual masses of LNG or spillets. The model is designed to predict the gas mass balance as a function of time and to display the spatial and temporal evolution of the gas (and radiated energy field).

LNGMAP has been validated by comparisons to predictions of models developed by ABS Consulting and Sandia for time dependent point releases from a draining tank, with and without burning. Simulations were in excellent agreement with those performed by ABS Consulting and consistent with Sandia's steady state results.

To illustrate the model predictive capability for realistic emergency scenarios, simulations were performed for a tanker entering Block Island Sound. Three hypothetical cases were studied: the first assumes the vessel continues on course after the spill starts, the second that the vessel stops as soon as practical after the release begins (3 min), and the third that the vessel grounds at the closest site practical. The model shows that the areas of the surface pool and the incident thermal radiation field (with burning) are minimized and dispersed vapor cloud area (without burning) maximized if the vessel continues on course. For this case the surface pool area, with burning, is substantially smaller than for the without burning case because of the higher mass loss rate from the surface pool due to burning. Since the vessel speed substantially exceeds the spill spreading rate, both the thermal radiation fields and surface pool trail the vessel. The relative directions and speeds of the wind and vessel movement govern the orientation of the dispersed plume.

If the vessel stops, the areas of the surface pool and incident radiation field (with burning) are maximized and the dispersed cloud area (without burning) minimized. The longer the delay in stopping the vessel, the smaller the peak values are for the pool area and the size of the thermal radiation field. Once the vessel stops, the spill pool is adjacent to the vessel and moving down current. The thermal radiation field is oriented similarly. These results may be particularly useful in contingency planning for underway vessels.     

Analysis of large LNG spills on water part 1: Liquid spread and evaporation

The first part of this two-part review considers the theoretical and experimental results obtained on liquid spread and evaporation on large LNG spills on water. Both instantaneous spills, in which the spill time is much smaller than the time for complete vaporization, and continuous spills are considered. Also applications of the correlations are discussed.     

Application of CFD (Fluent) to LNG spills into geometrically complex environments

Recent discussions on the fate of LNG spills into impoundments have suggested that the commonly used combination of SOURCE5 and DEGADIS to predict the flammable vapor dispersion distances is not accurate, as it does not account for vapor entrainment by wind. SOURCE5 assumes the vapor layer to grow upward uniformly in the form of a quiescent saturated gas cloud that ultimately spills over impoundment walls. The rate of spillage is then used as the source term for DEGADIS. A more rigorous approach to predict the flammable vapor dispersion distance is to use a computational fluid dynamics (CFD) model. CFD codes can take into account the physical phenomena that govern the fate of LNG spills into impoundments, such as the mixing between air and the evaporated gas. Before a CFD code can be proposed as an alternate method for the prediction of flammable vapor cloud distances, it has to be validated with proper experimental data. This paper describes the use of Fluent, a widely-used commercial CFD code, to simulate one of the tests in the "Falcon" series of LNG spill tests. The "Falcon" test series was the only series that specifically addressed the effects of impoundment walls and construction obstructions on the behavior and dispersion of the vapor cloud. Most other tests, such as the Coyote and the Burro series, involved spills onto water and relatively flat ground. The paper discusses the critical parameters necessary for a CFD model to accurately predict the behavior of a cryogenic spill in a geometrically complex domain, and presents comparisons between the gas concentrations measured during the Falcon-1 test and those predicted using Fluent. Finally, the paper discusses the effect vapor barriers have in containing part of the spill thereby shortening the ignitable vapor cloud and therefore the required hazard area. This issue was addressed by comparing the Falcon-1 simulation (spill into the impoundment) with the simulation of an identical spill without any impoundment walls, or obstacles within the impoundment area.     

Model of large pool fires

A two zone entrainment model of pool fires is proposed to depict the fluid flow and flame properties of the fire. Consisting of combustion and plume zones, it provides a consistent scheme for developing non-dimensional scaling parameters for correlating and extrapolating pool fire visible flame length, flame tilt, surface emissive power, and fuel evaporation rate. The model is extended to include grey gas thermal radiation from soot particles in the flame zone, accounting for emission and absorption in both optically thin and thick regions. A model of convective heat transfer from the combustion zone to the liquid fuel pool, and from a water substrate to cryogenic fuel pools spreading on water, provides evaporation rates for both adiabatic and non-adiabatic fires. The model is tested against field measurements of large scale pool fires, principally of LNG, and is generally in agreement with experimental values of all variables.     

LNG fires: a review of experimental results, models and hazard prediction challenges

A number of experimental investigations of LNG fires (of sizes 35 m diameter and smaller) were undertaken, world wide, during the 1970s and 1980s to study their physical and radiative characteristics. This paper reviews the published data from several of these tests including from the largest test to date, the 35 m, Montoir tests.

Also reviewed in this paper is the state of the art in modeling LNG pool and vapor fires, including thermal radiation hazard modeling. The review is limited to considering the integral and semi-empirical models (solid flame and point source); CFD models are not reviewed. Several aspects of modeling LNG fires are reviewed including, the physical characteristics, such as the (visible) fire size and shape, tilt and drag in windy conditions, smoke production, radiant thermal output, etc., and the consideration of experimental data in the models. Comparisons of model results with experimental data are indicated and current deficiencies in modeling are discussed.

The requirements in the US and European regulations related to LNG fire hazard assessment are reviewed, in brief, in the light of model inaccuracies, criteria for hazards to people and structures, and the effects of mitigating circumstances. The paper identifies: (i) critical parameters for which there exist no data, (ii) uncertainties and unknowns in modeling and (iii) deficiencies and gaps in current regulatory recipes for predicting hazards.     

Heat radiation from fires of organic peroxides as compared with propellant fires

This work was carried out to provide information on the safe distance of a store containing organic peroxides from its surroundings. Burning tests were     

常压粉末绝热贮槽及槽车汽化氧气的回收

介绍了将液氧在生产贮存过程中汽化的氧气 ,以及槽车充液前卸压放空的氧气回收充瓶的流程。该方法投资少 ,收益快 ,实际运行安全、可靠     

Comparison of spectra using a Bayesian approach. An argument using oil spills as an example

The problem of assigning a probability of matching a number of spectra is addressed. The context is in environmental spills when an EPA needs to show that the material from a polluting spill (e.g., oil) is likely to have originated at a particular site (factory, refinery) or from a vehicle (road tanker or ship). Samples are taken from the spill, and candidate sources and are analyzed by spectroscopy (IR, fluorescence) or chromatography (GC or GC/MS). A matching algorithm is applied to pairs of spectra giving a single statistic (R). This can be a point-to-point match giving a correlation coefficient or a Euclidean distance or a derivative of these parameters. The distributions of R for same and different samples are established from existing data. For matching statistics with values in the range {0,1} corresponding to no match (0) to a perfect match (1) a beta distribution can be fitted to most data. The values of R from the match of the spectrum of a spilled oil and of each of a number of suspects are calculated and Bayes' theorem is applied to give a probability of matches between spill sample and each candidate and the probability of no match at all. The method is most effective when simple inspection of the matching parameters does not lead to an obvious conclusion; i.e., there is overlap of the distributions giving rise to dubiety of an assignment. The probability of finding a matching statistic if there were a match to the probability of finding it if there were no match, expressed as a ratio (called the likelihood ratio), is a sensitive and useful parameter to guide the analyst. It is proposed that this approach may be acceptable to a court of law and avoid challenges of apparently subjective opinion of an analyst. Examples of matching the fluorescence and infrared spectra of diesel oils are given.