Analysis of domino effect in pipelines.

Parallel pipelines are frequently installed over long distances, due to the difficulty in creating or maintaining the required corridor. This implies that a release in one pipeline can seriously affect another one. The main risks associated with this domino effect are erosion by fluid-sand jets and the thermal action of jet fires. In this paper a survey has been performed on the accidents that have occurred, and the diverse associated domino sequences are analyzed. The probability of occurrence of domino effect is a function of the location of the hole, the jet direction and solid angle, the diameter of both pipelines and the distance between them. A mathematical model has been developed to estimate this probability. The model shows how the probability of domino effect decreases with the distance and diameter of the source pipe, and increases with the diameter of the target pipe. Its frequency can be estimated from this probability and from the frequency of the initiating pipe failure plus, in the case of jet fire impingement, the probability of ignition. The frequency of the target pipe failure thus calculated, always higher than its individual frequency, allows a more realistic risk analysis.

A dispersion safety factor for LNG vapor clouds.

The growing importance of liquefied natural gas (LNG) to global energy demand has increased interest in the possible hazards associated with its storage and transportation. Concerning the event of an LNG spill, a study was performed on the relationship between the distance at which the lower flammability limit (LFL) concentration occurs and that corresponding to the visible contour of LNG vapor clouds. A parameter called the dispersion safety factor (DSF) has been defined as the ratio between these two lengths, and two expressions are proposed to estimate it. During an emergency, the DSF can be a helpful parameter to indicate the danger of cloud ignition and flash fire.

Convective heat transfer around vertical jet fires: an experimental study.

The convection heat transfer phenomenon in vertical jet fires was experimentally analyzed. In these experiments, turbulent propane flames were generated in subsonic as well as sonic regimes. The experimental data demonstrated that the rate of convection heat transfer increases by increasing the length of the flame. Assuming the solid flame model, the convection heat transfer coefficient was calculated. Two equations in terms of adimensional numbers were developed. It was found out that the Nusselt number attains greater values for higher values of the Rayleigh and Reynolds numbers. On the other hand, the Froude number was analyzed only for the subsonic flames where the Nusselt number grows by this number and the diameter of the orifice.

Domino effect in chemical accidents: main features and accident sequences.

The main features of domino accidents in process/storage plants and in the transportation of hazardous materials were studied through an analysis of 225 accidents involving this effect. Data on these accidents, which occurred after 1961, were taken from several sources. Aspects analyzed included the accident scenario, the type of accident, the materials involved, the causes and consequences and the most common accident sequences. The analysis showed that the most frequent causes are external events (31%) and mechanical failure (29%). Storage areas (35%) and process plants (28%) are by far the most common settings for domino accidents. Eighty-nine per cent of the accidents involved flammable materials, the most frequent of which was LPG. The domino effect sequences were analyzed using relative probability event trees. The most frequent sequences were explosion→fire (27.6%), fire→explosion (27.5%) and fire→fire (17.8%).

Axial temperature distribution in vertical jet fires.

The behaviour of vertical commercial propane jet fires (flame lengths of up to 8m) was studied experimentally. The temperatures along the jet fire centreline were measured using a set of thermocouples and the flame contour was determined from infra-red (IR) images. The results show that temperature increases from the bottom of the flame, reaches a maximum value and decreases again at the top zone. A second-degree polynomial expression describes fairly well the variation of temperature as a function of the position on the flame centreline. The temperature along the centreline was found to increase for Q values lower than 7MW and to decrease at higher values.

Predicting the emissive power of hydrocarbon pool fires.

The emissive power (E) of a flame depends on the size of the fire and the type of fuel. In fact, it changes significantly over the flame surface: the zones of luminous flame have high emittance, while those covered by smoke have low E values. The emissive power of each zone (that is, the luminous or clear flame and the non-luminous or smoky flame) and the portion of total flame area they occupy must be assessed when a two-zone model is used. In this study, data obtained from an experimental set-up were used to estimate the emissive power of fires and its behaviour as a function of pool size. The experiments were performed using gasoline and diesel oil as fuel. Five concentric circular pools (1.5, 3, 4, 5 and 6m in diameter) were used. Appropriate instruments were employed to determine the main features of the fires. By superimposing IR and VHS images it was possible to accurately identify the luminous and non-luminous zones of the fire. Mathematical expressions were obtained that give a more accurate prediction of E(lum), E(soot) and the average emissive power of a fire as a function of its luminous and smoky zones. These expressions can be used in a two-zone model to obtain a better prediction of the thermal radiation. The value of the radiative fraction was determined from the thermal flux measured with radiometers. An expression is also proposed for estimating the radiative fraction.

Experimental study of thin-layer boilover in large-scale pool fires.

By performing a series of pool fire experiments, the authors attempt to apply knowledge of thin-layer boilover to the large scale (pool diameter from 1.5 to 6 m). Two commercial hydrocarbons were used: gasoline and diesel. As expected, only in the case of diesel did the phenomenon of thin-layer boilover occur. Data were used to analyze various features of thin-layer boilover in fires, particularly its intensity and onset time. A comparison with results published in the literature shows the importance of this study.

Using liquid superheating energy for a quick estimation of overpressure in BLEVEs and similar explosions.

A method is proposed for the quick estimation of the peak overpressure caused by a Boiling Liquid Expanding Vapor Explosion (BLEVE) or a similar explosion. The method is based on the use of the "superheating energy" (SE), which is the difference between the specific enthalpy of the liquid at the temperature just before the explosion and the specific enthalpy of the liquid at its saturation temperature, at atmospheric pressure. The analysis performed with a set of reference substances showed that in a BLEVE or in similar explosions, the energy converted into overpressure will range between 3.5 and 14% of SE. The comparison of the values thus obtained with experimental data from the literature shows a fairly good agreement.

Consequences of major accidents: assessing the number of injured people.

Quantitative risk assessment studies of accident scenarios usually involve estimating the number of fatalities that can be expected. The number of people injured, however, is seldom evaluated because it implies significant additional effort and often the information required to perform this evaluation is not available. However, the number of injured people can be very important for emergency planning, especially in relatively large accidents. In this paper, a set of 975 accidents were selected for analysis, with the aim of searching for a relationship between the number of people killed and the number of people injured. As the data were scattered, principal component analysis and clustering analysis were applied to identify the data subsets that could undergo a selective, specific statistical treatment. Further treatment of these subsets led to mathematical expressions that are used to estimate the probable number of injured people as a function of the number of fatalities for all accidents, as well as for gas cloud, fire and explosion events, respectively.