Showing papers in "Journal of Loss Prevention in The Process Industries in 1990"

Sizing of Dust Explosion Vents in the Process Industries: Further Consideration of Some Important Aspects

Abstract: This chapter provides an overview of some vent sizing methods used in Europe and the United States.. The vent ratio is defined as vent area per unit of enclosure volume. Originally, a fixed vent ratio, determined by the maximum rate of pressure rise in the 1.2 liter Hartmann bomb, was specified for a specific dust. However, as the enclosure volume gets larger, the required vent area increases. The nomograph method was originally designed by Verein deutscher Ingenieure (1979), but later it was also adopted by the National Fire Protection Association in the United States. The relevance of this method is tied to the nature of the large-scale experiments on which it rests. These experiments were conducted with dust clouds generated by blowing dust into the experimental enclosures from pressurized reservoirs through narrow nozzles to ensure uniform, well-dispersed, and highly turbulent dust clouds. . The Norwegian method and the Randandt scaling law are also discussed in this chapter. It is necessary, in each specific case, to analyze carefully what kind of ignition sources are likely to occur and what locations within the silo volume ignition have a significant probability. The theories by Maisey, Heinrich and Kowall, Rust, Nomura and Tanaka, Nagy and Verakis, Gruber et al, Swift, and Ural are explained in the chapter. The venting philosophy outlined in the chapter implies that even a modestly sized vent may add significantly to the safety standard of the plant by being capable of providing adequate relief for the majority of the expected explosions.

Fast flame speeds and rates of pressure rise in the initial period of gas explosions in large L/D cylindrical enclosures

Abstract: Flame speeds and rates of pressure rise for gaseous explosions in a 76 mm diameter closed cylindrical vessel of large length to diameter ratio ( L/D = 21.6), were quantitatively investigated. Methane, propane, ethylene and hydrogen mixtures with air were studied across their respective flammability ranges. Ignition was affected at one end of the vessel. Very fast flame speeds corresponding to high rates of pressure rise were measured in the initial 5–10% of the total explosion time. During this period 20–35% of the maximum explosion pressure was produced, and over half of the flame propagation distance was completed. Previous work has concentrated on the later stages of this type of explosion; the development of tulip flames, pressure wave effects and transition to turbulence. The initial fast phase is very important and should dominate considerations in pressure relief vent design for vessels of large L/D .

Severity of Accidents with Hazardous Materials. A Comparison between Transportation and Fixed Installations

Abstract: A review of accident case histories relevant to hazardous materials has been performed. From the literature, 1793 accident case histories involving hazardous materials were identified, most of them from the period 1960–1988. Of these 1793 accidents, 39% happened during transportation. In 682 accidents the consequences included fatalities, and of these 27% involved the transportation of hazardous materials. The accumulated frequency-fatality curves (so-called fN curves) have been constructed and are close to straight lines with a slope of –1, indicating that the probability of having an accident with, for instance, more than one hundred fatalities is approximately ten times lower than the probability of having an accident with more than ten fatalities. The accidents were grouped according to transportation type, and the difference between the various groups was tested using a standard χ 2 approach. No significant difference between fixed installations and transportation was found for accidents having consequences above three fatalities. Differences were found between Western Europe and North America and the rest of the world. Once an accident has happened, it seems to have more severe effects outside Western Europe and North America, both for transportation accidents and fixed installations. Comparison of fN curves for Western Europe and North America from different time periods did not show significant differences, indicating that once an accident has happened, it has the same probability of escalating now, as it had 20 years ago.

A comprehensive program for calculation of flame radiation levels

Abstract: A comprehensive program for the calculation of radiation levels from a jet flame, pool fire or BLEVE will be described. The program initially calculates the flame shapes, and may then be used to calculate the radiation level at any specified point relative to the flame, or the contour of a specified radiation level. The state-of-the-art models that the program employs are discussed. An assessment of the risk from flammable effects through the use of probit relations will be made.

Two-phase flow discharge in nozzles and pipes — a unified approach☆

Abstract: This paper presents a unified treatment on homogeneous two-phase discharge through short nozzles and long pipes. The resulting generalized solutions, cast in terms of a limited number of dimensionless variables, can adequately account for liquid properties, inlet conditions, effect due to pipe friction, gravitational change, inlet subcooling and presence of non-condensable gases. The results are also presented in the form of design charts, for practising engineers.

Rapid depressurization of pressure vessels

Abstract: Experiments were conducted on the rapid depressurization of large pressure vessels. Measurements taken included the pressure, temperatures at a large number of positions both within the fluid phase(s) and on the wall of the vessel, and composition, all as a function of time during the blowdown process. The systems studied included subcritical and supercritical, condensing and non-condensing. From these experiments, an understanding of the physical processes involved during blowdown was evolved. This was incorporated into a mathematical model of blowdown, and implemented in a computer program. The model correctly predicts all the major phenomena observed in the experiments, as a function of time.

A new general method for estimation of heats of combustion for hazard evaluation

Abstract: A new method is presented for estimating heats of combustion. Existing methods are restricted to estimating values for compounds composed of a small set of elements. The method presented is much more general, being capable of providing values for compounds or mixtures of compounds composed of any of 71 elements. Errors for tested organic, organometallic, or inorganic compounds were typically less than a few per cent. The new method involves expert based heuristics that are extended by thermodynamic approximation. The method is suitable for computer implementation in applications such as prediction of incinerator operations or hazard evaluations.

Discharge coefficients for choked gas—liquid flow through nozzles and orifices and applications to safety devices

Abstract: A choked flow model is employed to analyse the available experimental data on compressible gas-liquid flow through nozzles and sharp-edged orifice plates. The data satisfying the conditions for choked flow are subjected to further analysis to determine correlations for the discharge coefficients. It is demonstrated that choking mass flowrates can be predicted to within ±6% for nozzles and ±7% for sharp-edged orifice plates. These correlations for discharge coefficients are intended to be of a general nature and ought to lend some confidence to their applicability outside the ranges of experimental parameters. The applicability of these results to pressure relief valves and rupture discs is also discussed.

Comparison of an integral model for predicting the dispersion of a turbulent jet in a crossflow with experimental data

Abstract: The safe design of high pressure plant and pipework requires the installation of adequate pressure relief and vent stacks. To ensure that any discharge is safety removed, it is necessary to consider how the gas disperses under a variety of atmospheric conditions. This is best achieved by developing a mathematical model that has been validated against reliable data. This paper gives details of a mathematical model for predicting the dispersion of a jet in a crossflow. The model is based upon well-tried similarity ideas, and contains a number of closure or entrainment coefficients whose values must be specified. Appropriate well-documented laboratory or field-scale tests are selected to enable these values to be chosen. The predictions of the resulting model are than compared with experimental results over a wide range of conditions, including free and wind-blown underexpanded sonic and subsonic jets of natural gas and other materials.

The influence of flow blockage on the rate of pressure rise in Large L/D cylindrical closed vessel explosions

Abstract: A quantitative determination was made of the influence of the blockage of a single baffle on the increase in the rate of pressure rise in a large length/diameter (L/D) ratio closed vessel explosion. A 76 mm diameter vessel with an L/D of 2 was used, and a four hole flat grid plate was used to simulate the shear layer size and turbulence levels of practical obstructions of complex shape. The maximum rate of pressure rise was investigated as a function of the baffle blockage and position, relative to the spark at the base of the vessel. The unburned gas set in motion ahead of the flame was measured by determining the differential pressure across the obstacle. This showed that mean velocities ahead of the flame were in the range 15–20 m s−1 upstream of the baffle, and were caused by the high flame speed, with maximum values up to 30 m s−1. These high velocities caused large turbulence levels to be created on interaction with the grid plate, depending on the grid plate blockage. Flame acceleration occurred once the flame reached the turbulence, and this acceleration was shown to be a function of the blockage. An approximate method for relating the increase in the rate of pressure rise to turbulent burning velocity measurements was developed, using a simple method of estimating the turbulence levels generated downstream of the obstacle.

Sizing of dust explosion vents in the process industries: Advances made during the 1980s

Abstract: To prevent damage to humans and property caused by dust explosions in the process industries, several kinds of precautions may be taken. Because of its favourable features, both technically and economically, venting has become a widely used method for controlling dust explosions. It is therefore unfortunate that the literature on dust explosion vent sizing is still in part contradictory. It is generally appreciated that the vent area needed to keep the pressure in a dust explosion in a given enclosure below a given limit. is determined by the rate of heat release in the explosion. However, this rate depends not only on specific dust properties, such as chemistry and fineness, but also on several properties of the dust cloud that are determined entirely by the industrial process in which the cloud is generated and ignited. Taking these properties into account has become a central issue during the 1980s. This paper reviews a number of realistic large-scale experimental investigations conducted during recent years. The experimental results are compared with the corresponding vent areas recommended by several different vent sizing methods in current use. A differentiated method for vent sizing, based on data from experiments conducted under realistic industrial conditions, is proposed. The continued need for conducting further realistic experiments in various types of full-scale vented indusrial enclosures is emphasized.

Water spray mitigation of hydrofluoric acid releases

Abstract: Two series of tests were conducted in flow chambers to assess the effectiveness of water spray in mitigating releases of hydrogen fluoride (HF). Bench scale experiments identified key variables for testing in a larger facility. The larger scale field tests demonstrated that HF releases can be mitigated with water, and also quantified the impact of numerous design variables on mitigation effectiveness. HF removal efficiencies of 25–90% were demonstrated at water to HF liquid ratios of 6:1 to 40:1.

An integrated model for discharge rate, pool spread, and dispersion from punctured process vessels☆

Abstract: Extensions have been made to a number of previously-published models to create an integrated system for predicting the effects of an accidental release from a process vessels. Single or two-phase discharge rates are calculated by the LEAKR model developed for Environment Canada. This has been extended using technology developed for DIERS (Design Institute of Emergency Relief Systems) and DIPPR (Design Institute of Physical Properties Research). Aerosol formation and rain-out were modelled, followed by a full transient pool spread and vaporization model. Other extensions were made to the ONDEK model for elevated, high momentum releases and to the DEGADIS model for ground level releases. Comparisons are reported here of model predictions with experimental data for pool spreading and evaporation rate.

Two-phase flashing flow methods and comparisons

Abstract: Two-phase flashing flow methods used by DIERS, the American Petroleum Institute (API) and ASME are reviewed and compared. For saturated and low quality inlet conditions, the API method was found to yield higher theoretical mass flow rates than the DIERS homogeneous equilibrium flow model (HEM). The degree of disparity can be quantified by two dimensionless physical property groups. Comparison with flashing flow data for water, refrigerant 12 and a propane—butane mixture shows good agreement with the HEM prediction, but poor agreement with the API method. The ASME method only recommends a saturated water flashing flow curve that is in agreement with the HEM calculation to within 10%.

An overview of equipment for containment and disposal of emergency relief system effluents

Abstract: This paper presents an overview of the various types of equipment and systems used in the containment and disposal of effluents from emergency relief systems, i.e., safety valves and rupture discs. The paper discusses the design, sizing, safety considerations, mechanical aspects, and operational considerations for such equipment as containment/separation vessels (blowdown drums, knock-out drums, catchtanks and quench tanks) and disposal equipment such as vent stacks, flare stacks and systems, vent condensers and scrubbers (gas absorbers).

Pressure discharge from a pressure vessel filled with CO2

Abstract: Recently in high-pressure process technology, supercritical gas extraction has been carried out on a large scale. CO 2 is often used as compressed gas, particularly in food processing technology. Batch processes exsist that involve pressures of between 150 and 300 bar. On termination of the extraction process, the pressure vessels must be relieved. According to the thermodynamic equilibrium ratios, CO 2 temperatures drop with the formation of dry ice. Safety risks hence involve the possible icing-up and clogging of blow-off lines and relief valves, and the occurrence of thermal stresses on pressure vessels fitted with austenitic inner linings. Test results are shown that were obtained when relieving a 50 l pressure vessel with and without filling of the feedstock, the objective being a short relief time with the safest temperature for the vessel inner walls.

A model of the dilution of a forced two-phase chemical plume in a horizontal wind

Abstract: Accidendtal releases of volatile liquid chemicals from pressurized storage vessels result in the formation of high-momentum flashed jets with liquid phases comprised of extremely fine droplets (aerosol). These jets are typically heavier than air. The subject of this paper is the description of a turbulent entrainment model that predicts the trajectory and dilution of such releases when the release point is elevated above ground-level in a horizontal wind. The model is compared with available field test data for pressurized liquid releases and is found to describe the observations to a degree of accuracy adequate for most hazard assessment purposes. The comparisons include the field measurements made during the liquid ammonia (NH 3 ) and liquid hydrofluoric acid (HF) release test series, namely Desert Tortoise and Goldfish respectively, at the US Department of Energy Fuels Test Facility. The model indicates that over the 3000 m distance in which measurements were made, there is no need to invoke atmospheric turbulence to explain the dilution of the ammonia jet. However, atmospheric turbulence is predicted to dominate the mixing in the far field of the weaker HF jet.

Jets discharging to atmosphere

Abstract: Consideration is given to the range of jets that may arise from the safety or accidental venting of the contents of pressurized liquid or gas storage vessels. Attention is focussed on the momentum dominated region of the discharge, and the present analytical description of gaseous jets with a view to safety analysis is reviewed. Data are presented for a number of two-phase jets of superheated Freon 11, including temperatures, velocities, mass fractions and droplet sizes, and these are shown to be strongly influenced by the liquid/vapour non-equilibrium in the near field and the subsequent evaporating nature of the liquid phase and the accompanying air entrainment. The experimental needs in these complex two-phase systems are discussed with a view to obtaining a better understanding of the source terms. Some recent developments using laser fluorescence as a selective probe are presented. These are shown to have the potential to resolve some of the properties of the individual liquid and vapour phases. The case of liquid temperature is examined.

Influence of initial pressure on spark-ignited dust explosions

Abstract: An investigation was conducted on the effect of initial pressure on lycopodium dust explosions in a 26 I test chamber. Maximum explosion pressures and maximum rates of pressure rise were measured over a range of initial pressures from 0.87 bar to 2.0 bar. Spark ignition was used for all tests. For each initial pressure, the peak values of maximum explosion pressure and maximum rate of pressure rise were found to occur at a fuel equivalence ratio of 4.5. When considered as a function of initial pressure, the peak explosion pressure values displayed a linear and proportional variation.

Comparison of methods of hazard identification at plant level

Abstract: Plant level hazard identification methods (Dow index and Mond index) have been evaluated by analysis of three different chemical plants (a reforming plant, a fine chemicals plant and a fertilizer plant). The results showed good agreement between the hazard rankings found by use of the Dow and Mond indices, respectively. However, some discrepancies were found, e.g. with respect to hazard ranking of gaseous hydrogen. One important limitation of the Dow and Mond indices is the lack of toxicological assessment of the involved chemical substances that were found to cause an incorrect hazard ranking when a toxic substance such as NH 3 is handled. A number of improvements to the methods are suggested.

Effects of vapour/aerosol and pool formation on rupture of vessels containing superheated liquid

Abstract: Experiments have been conducted involving Freon 114 (boiling point 3.5°C at 1 bar), contained in spherical glass containers that were suddenly ruptured. The vapour/droplet cloud formed and droplet sizes have been measured with videos and high speed movies. Pool formation has been measured by collecting the liquid on a cold tray. The main parameters varied were the liquid superheat, the height of release and the percentage of container filling. The superheated liquid shattered into droplets at a vapour volume fraction of ≈50%. The velocity and size of the largest droplets had Weber numbers ≈20–30. The droplets velocities were, however, an order of magnitude smaller than predicted by isentropic expansion of the mixture down to atmospheric conditions—a model that is often used for catastrophic releases of liquefied gases. The droplet velocities were overpedicted by isentropic expansion from bubbly to slug flow. The overprediction was caused by entropy production, which has been associated with bubble nucleation and growth. The liquid pool appeared to form due to drop impingement on the ground. It contained only a fraction of the liquid that would have been deposited, based on adiabatic (or isentropic) flash fractions. A model for vapour cloud formation, based on the evaporation of the drops as they travel outwards appears consistent with these observations, and with those made using a smoke curtain.

Release conditions following loss of containment

Abstract: During the past decade a series of experiments has been carried out to determine the source term parameters of two-phase releases following the catastrophic failure of vessels containing superheated, low boiling point liquids. The aim of this work was to model the release of LPG and similar substances following failure of storage vessels. Various fracture mechanisms of the vessels were studied, and experimental data were obtained on pressure decay, fracture pattern, and on the resulting aerosol, vapour and residual liquid content of each release. Emphasis was placed on gaining data on the overall cloud dynamics and on aerosol properties, such as particle size distribution, velocity and concentration. Such experimental data were used to examine the applicability of the few existing models relating to source term characteristics, and to provide information necessary for a complete thermodynamic appreciation of the events that occur immediately before, during and after release.

The effects of sloping terrain on dense gas dispersion

Abstract: Although many well-validated models exist for the prediction of dense gas dispersion, in most cases these have been developed for idealized flat terrain conditions. When applying such models to safety case studies, care must be exercised to ensure that the results are realistic even for non-idealized situations. A standard flat-terrain dispersion model has therefore been developed to incorporate the effects of significant slopes. Certain categories of release undergo considerable acceleration down hills, resulting in greater hazard ranges, whereas other types of release are hardly affected. The presence of significant slopes near a potential release source can therefore significantly modify the consequences analysis results incorporated in a safety case.

Critical two-phase mass flow rates of liquefied gases

Abstract: The critical mass flow rates were calculated for a wide variety of substances, including propane, ethylene, ammonia, chlorine, vinyl chloride, water and refrigerants. The mass flow calculations were based on isentropic homogeneous flow in thermal equilibrium and on newly generated thermodynamic data derived from appropriate equations of state such as the Benedict-Webb- Rubin equation, Starling's equation and the Martin-Hou equation. A generalized equation was developed to express the critical mass flux in a simple algebraic expression as a function of temperature, pressure and molecular weight. This universal two-phase flow prediction equation gave good results on all substances tested.

Thermal radiation hazards from organic peroxides

Abstract: This paper reviews the approaches used in FRG, the Netherlands and by Laporte Industries in the UK to assess the thermal radiation hazards from organic peroxides and derive quantity safety distances relating to them. The evidence suggests that the actual behaviour of packaged material cannot be reliably determined by either calculation or small scale laboratory methods. A possible system, utilizing package trial data, is presented to predict the variation of irradiance with distance.

A theoretical study of absorption of toxic gases by spraying

Abstract: The feasibility of controlling toxic gases by absorption using liquid sprays has been examined previously. In this paper the physical and chemical phenomena taking place in a gas-spray environment are described in more detail, and a mathematical model of the interactions is developed. The model was implemented in the PSI-Cell computer code and numerical solutions were obtained for a number of different configurations and spray patterns (i.e. vertical and horizontal co-current and counter-current flow, upwards and downwards cross-flow). The effects of several spray parameters (e.g. flow rate, location and drop-size distribution) and gas parameters (e.g. solubility, chemical reaction rate) were considered. Several simulations of water spraying of specific gases (HF, NH3, SO2 and H2S) were carried out. The effectiveness of gas removal estimated by the numerical model was found to compare favourably with HF laboratory tests. Favourable agreement was also shown with qualitative field data on NH3 and SO2 spraying. Absorption using water spraying provided an effective means of control for highly water soluble gases with fast ionization reactions in the liquid phase (e.g. HF, NH3), but it did not result in substantial gas removal when the gases were only moderately water soluble (e.g. SO2).

Sulphur trioxide and oleum hazard assessment

Abstract: Sulphur trioxide and strong oleums are some of the most corrosive materials used in general chemical manufacture. Because of their explosive reactivity with water and water vapour, the evaluation of hazard and the consequences to humans following a spill is difficult and no published method exists. Following amendment of the EC Directive, sulphur trioxide has now been added to the list of substances under the CIMAH regulations for which installations will have to prepare safety reports for submission to HSE. This paper reviews past incidents involving these materials and gives a method for assessing hazard and consequences dealing with spills on wet or dry ground. Using a toxicology criterion, a method for calculating maximum downwind hazard ranges is given for a range of typical release cases. The effectiveness of certain emergency actions and safeguards are discussed.

Atmospheric release tests of monomethylamine

Abstract: Reduced storage temperature is one preventive measure that can be employed to lessen the consequences from accidental releases. A literature survey indicates that a large reduction in potential cloud formation can be obtained by storing a material below the point at which vapour flashing, due to liquid superheat, will completely atomize the release. To extend and confirm these studies to monomethylamine, 10 atmospheric release were made using a 6.3 mm (0.25 inch) orifice. The results indicate that flash atomization can be prevented by storing monomethylamine with