Contents
A vapour cloud explosion is an explosion arising when a flammable gas or vapour cloud ignites and its combustion produces a blast wave. Zalosh describes large clouds burning among densely arranged process equipment as an important example. The resulting blast depends on flame speed and confinement, as well as the combustible material present. The released fuel mass alone therefore does not determine the pressure imposed on surrounding equipment or structures.[1]
The assessment connects a release, mixing with air, ignition, flame propagation and pressure loading. Each stage can change the location and extent of the hazard. A cloud may remain outside, enter a building through ventilation openings, or spread between interconnected compartments. The hydrogen scenarios in the EPRI/NRC methodology illustrate these alternative pathways within US nuclear power facilities.[2]
Formation and ignition of a cloud
A release does not necessarily ignite at its source. Fuel can mix with air before encountering an ignition source elsewhere. The resulting concentration field depends on the release and its dispersion. For outdoor hydrogen releases, the EPRI/NRC methodology considers rising gas, wind direction and atmospheric conditions when examining whether a cloud reaches an ignition location.[2]
The flammability limits identify concentration conditions within which flame propagation is possible under specified test conditions. They do not define the shape of a real cloud. Concentration can vary between locations, and the entire released inventory need not belong to one combustible mixture. Zalosh also distinguishes concentration limits from the ignition energy needed to initiate combustion.[3]
The distinction between immediate and delayed ignition changes the scenario. The EPRI/NRC report treats immediate ignition of a pressurised hydrogen leak as a jet flame. It separately considers gas accumulation followed by ignition and possible return of combustion towards a continuing release. These branches require different exposure descriptions because the combustible material occupies different locations when burning begins.[4]
Release duration and the time to ignition therefore influence the material available for combustion. In the report's analytical approach, the time for gas to reach an ignition source can also be compared with possible isolation actions. That comparison does not establish an automatic successful intervention. Its probability belongs to the scenario assessment and requires evidence about the relevant equipment and actions.[2]
Flame propagation, congestion and confinement
An explosion assessment must distinguish the fuel's chemical energy from the rate and geometry of its release. A deflagration propagates through the unburned mixture at subsonic speed. Expansion of the hot burned gases drives motion in the surrounding mixture. The flame speed seen by a stationary observer consequently differs from the burning velocity relative to the unburned gas.[5]
Turbulence can increase burning through faster local propagation or a more distorted flame with greater surface area. This helps explain why a cloud among obstructions cannot be represented by combustible mass alone. The handbook associates vapour cloud blast strength with the flame speed that develops and the degree of confinement. The equipment arrangement is therefore part of the physical problem.[5][1]
Congestion and confinement describe related but different features of that arrangement. The handbook's example combines a high density of process equipment with restrictions in and around the cloud. An outdoor process area can contain extensive obstructions even without forming a closed room. Conversely, gas entering a building can encounter a substantially different boundary arrangement from the outdoor release location.[1][2]
A cloud explosion is not automatically a detonation. A detonation involves supersonic propagation through the unburned mixture and a strongly nonuniform transient pressure field. The possibility of deflagration-to-detonation transition depends on mixture reactivity and the geometry available for flame acceleration. The handbook also identifies roughness and obstructions as relevant influences in elongated enclosures.[6]
These distinctions affect protection and structural assessment. A slowly developing enclosure pressure and a rapidly travelling pressure wave do not impose the same loading history. Zalosh therefore treats detonation protection separately from closed-vessel deflagration protection. A method selected for one propagation regime cannot be assumed suitable for the other simply because the fuel is identical.[6]
Pressure, impulse and ideal blast scaling
Blast pressure varies with time at a receiving location. The incident pressure describes the arriving wave, while reflection at a surface can alter the pressure applied there. The duration of the positive pressure phase also matters. Its accumulated effect, called impulse, is relevant to dynamic structural response.[7]
Ideal blast theory simplifies the source to energy released instantaneously at one point. The handbook defines a dimensionless radius by combining actual distance, ambient pressure and blast energy. This provides a useful way to explain the scaling assumption, although it does not establish its validity for a particular vapour cloud.[7]
| Symbol | Meaning | Unit |
|---|---|---|
| R̄ | Dimensionless radius in ideal blast scaling | dimensionless |
| R | Distance from the assumed point source | m |
| P0 | Ambient absolute pressure | Pa |
| E | Assumed energy contributing to the blast | J |
The pressure-to-energy ratio has units of inverse cubic metres. Its cube root therefore cancels the length dimension of the distance. At unchanged ambient pressure, maintaining the same scaled radius requires distance to vary with the cube root of blast energy. This is a mathematical consequence of the stated expression, not a separate prediction of cloud combustion.[7]
For illustration, select a distance of 50 m, ambient pressure of 101,325 Pa and assumed blast energy of 100 MJ. The resulting dimensionless radius is approximately 5.02. Doubling the distance to 100 m gives the same scaled radius only if energy increases eightfold, to 800 MJ. These selected values demonstrate scaling and do not describe a measured accident.[7]
No pressure or safe separation distance follows from that calculation alone. It still requires an applicable pressure relationship and a justified blast-energy input. The chemical energy of released fuel is not automatically equal to the energy contributing to the blast. The handbook identifies uncertain energy yield as one weakness of early vapour cloud assessments.[1]
Why a single equivalent energy is insufficient
TNT equivalence is relevant when an explosion is represented through an equivalent compact-explosive source. The handbook explains that ideal blast correlations draw on theory and tests involving compact condensed-phase explosives. Their assumed instantaneous point release differs from combustion developing through an extended vapour cloud. Matching an energy number does not remove that difference.[7][1]
Zalosh reports that early ideal-blast assessments were unsuccessful for two related reasons. Much of the practical interest concerned locations near the cloud, where a point-source representation was inadequate. Even farther away, the reported pressure and impulse decayed differently from ideal blast predictions. An arbitrary yield adjustment consequently could not correct every part of the pressure field.[1]
The 2002 handbook describes two alternative modelling approaches. One estimates flame speed empirically and uses calculations of nonideal wave generation. The other represents two-dimensional or three-dimensional flame propagation and calculates acceleration as part of the solution. This is a description of the approaches discussed in that edition, rather than an inventory of present-day software.[1]
Both approaches require more information than total released mass. The flame-speed representation and geometry must correspond to the scenario being examined. The EPRI/NRC methodology likewise leaves tool selection and suitability to the analyst; its hydrogen appendix does not verify or validate a particular explosion model. A named computational method is therefore not itself evidence of adequate validation.[1][8]
Hydrogen migration and connected spaces
Hydrogen safety illustrates why the release location and explosion location need separate treatment. The EPRI/NRC report considers an outdoor cloud reaching a building's ventilation intake. The relevant assessment then includes concentration at the intake, mixing within the ventilation system and possible ignition locations. Ignition outside the building and ignition inside it are distinct scenarios.[9]
The same report considers migration through ceiling openings, wall openings and ventilation ducts between compartments. A release in one space can therefore require examination of connected spaces. The report separates explosions confined to the compartment of origin from those involving several compartments. Its conservative damage assumptions were developed for nuclear-plant risk assessment and are not universal physical damage laws.[10]
For connected compartments, the proposed analysis examines the quantity needed to create combustible mixtures and the time required to release that quantity. It then considers failure to isolate the leak within the available period. This sequence links accumulation with system response. It avoids treating the presence of a gas source as proof that every connected room contains the same mixture simultaneously.[10]
Interpretation and evidence limits
A consequence assessment needs a consistent chain from source conditions to target loading. The release and dispersion define where combustible mixtures can occur. Ignition and flame propagation define how energy is released. Pressure histories and the response of the exposed equipment then determine the consequence being assessed.[2][7][1]
The scope of supporting evidence remains attached to that chain. The EPRI/NRC appendix uses historical nuclear-industry events to support its scenario rules. Those rules do not provide a calibrated general-purpose overpressure model for every industrial site. Its explicit separation between scenario development and model verification preserves that distinction.[11]
An output expressed only as peak pressure also leaves part of the loading unspecified. Reflection, duration and impulse can influence structural response. Likewise, a quoted distance requires a defined source geometry and consequence criterion. The ideal scaling example demonstrates one relationship between inputs; it does not supply those missing physical and response assumptions.[7][1]
See also
Notes
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Zalosh (2002), p. 3-420.
- ↑ 2.0 2.1 2.2 2.3 2.4 Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-8–N-11.
- ↑ Zalosh (2002), pp. 3-402–3-405.
- ↑ Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-7–N-8.
- ↑ 5.0 5.1 Zalosh (2002), pp. 3-407–3-410.
- ↑ 6.0 6.1 Zalosh (2002), pp. 3-410–3-412.
- ↑ 7.0 7.1 7.2 7.3 7.4 7.5 7.6 Zalosh (2002), p. 3-419.
- ↑ Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), p. N-11.
- ↑ Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), p. N-9.
- ↑ 10.0 10.1 Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-10–N-11.
- ↑ Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-1 and N-11.
References
- Electric Power Research Institute, & U.S. Nuclear Regulatory Commission. (2005). EPRI/NRC-RES fire PRA methodology for nuclear power facilities: Volume 2, Detailed methodology (EPRI 1011989; NUREG/CR-6850).
- Zalosh, R. (2002). Explosion protection. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-402–3-421). National Fire Protection Association.