BLEVE

A BLEVE (boiling liquid expanding vapour explosion) is the sudden failure of a container holding a liquid above its normal boiling point, which then flashes into vapour while the container breaks apart. Many substances that are gases at ambient conditions, such as liquefied petroleum gas, are stored under pressure as a saturated mixture of liquid and vapour. A rupture of such a vessel releases the stored energy violently. The typical cause is a fire that heats the part of the shell above the liquid until the metal can no longer withstand the internal pressure.

A BLEVE threatens its surroundings in three ways: a pressure wave, fragments of the vessel that can be thrown considerable distances, and, if the contents are flammable, a short but very intense fireball. In quantitative risk assessment of process plants it is treated as a delayed event, typically following 8 to 15 minutes of fire exposure. The radiation from the fireball can be estimated from the mass of fuel with a simple empirical relation, and the result is compared with damage thresholds for people and equipment.

Definition and terminology

The acronym stands for "Boiling Liquid, Expanding Vapor Explosion", written with the American spelling in the source documents. NUREG-1805 quotes the definition of the National Fire Protection Association (NFPA): the failure of a major container into two or more pieces. The failure occurs at a moment when the contained liquid is above its boiling point at normal atmospheric pressure. The glossary of the same report, citing Nolan (2000), describes it as a catastrophic rupture of a pressurised vessel with the simultaneous ignition of the vaporising fluid. During the rupture a pressure wave may be produced, and a short, intense fireball occurs if the liquid is flammable.

The two definitions differ in one respect. The NFPA wording centres on the failure of the container and the state of the liquid, and does not require ignition. The glossary wording includes ignition and the fireball that follows when the liquid is flammable. Barry uses the term "BLEVE fireball" for this outcome. Zalosh notes that a rising fireball occurs in most BLEVEs of liquefied fuel gases exposed to severe fire.

The failure itself fits the class of physical explosions in NUREG-1805, in which the high pressure is generated by mechanical means without any chemical change. Its two examples are the external heating of a tank, which raises the internal pressure until the tank fails, and the sudden release of a superheated liquid that flash-evaporates. A BLEVE of a fire-exposed tank combines both. The burning of the released vapour afterwards is a separate, chemical process. Introduction to Fire Safety Engineering makes the same distinction for ordinary fuels: evaporation is a physical change, and only the vapour formed can burn.

How a BLEVE develops

In the vessel, the liquid and the vapour above it are stored under pressure in a saturated state. External heating of the tank increases the internal pressure. Evaporation takes energy from the surroundings. The contained liquid therefore acts as a heat absorber, so the wetted part of the shell usually stays relatively cool. Only the surfaces around the internal vapour space are at risk.

A liquefied gas vessel exposed to fire: the liquid cools the wetted shell, while the shell around the vapour space heats up and weakens. Based on Iqbal and Salley (2004), p. 5-15.

NUREG-1805, citing Nolan (1996), states that a BLEVE typically occurs in a metal container that has been overheated above 538 °C (1,000 °F). At that temperature the metal may no longer withstand the internal stress, and failure occurs. Most BLEVEs occur when containers are less than one half to one third full of liquid, that is, when a large part of the shell is not wetted.

Barry describes the sequence as it is used in risk studies. A primary fire overheats the pressurised vessel; this raises the internal pressure and weakens the shell, until the shell bursts and releases its contents as a large and very intense fireball. The BLEVE is generally considered a delayed event, with 8 to 15 minutes of fire exposure before failure. Among the possible exposing fires, an impinging jet fire is especially severe, far more so than a pool fire. Large-scale tests measured heat fluxes of up to 250 kW/m² for two-phase propane jets and 300 kW/m² for sonic natural gas jets. On fully engulfed cylindrical targets the average heat flux was 150 kW/m² for two-phase liquefied petroleum gas (LPG) jets and 200 kW/m² for natural gas jets.

Typical BLEVE sequence: a primary fire heats the vessel for several minutes until the shell bursts; the flashing liquid then causes a blast, fragments and, for flammable contents, a fireball. Based on Barry (2002), p. 5-184, and Iqbal and Salley (2004), pp. 5-15 and F-8.

Fire is not the only route to failure. A container can be damaged by impact, after which a crack grows because of the internal pressure, brittleness of the vessel material, or both. The metal can also be weakened by corrosion, internal overheating or manufacturing defects, until large cracks form or the vessel separates into two or more pieces.

When the shell of a vessel with a pressurised cargo opens, part of the released liquid flash-evaporates and a vapour cloud forms suddenly. The phase change produces blast waves whose energy equals the change in internal energy of the liquid and vapour. If the vapour puff is ignited immediately after it forms, it burns as a rising sphere, the fireball.

Hazards: blast, fragments and fireball

The blast wave from a BLEVE is one example of a breached-vessel blast. For a vessel holding a gas that does not ignite, ideal blast wave correlations based on the expansion energy of the gas can be used. BLEVEs of liquefied gases need their own correlations for blast and fragments. Zalosh refers to the guidelines of the Center for Chemical Process Safety for these correlations; they are not reproduced in the SFPE handbook.

Fragments are a hazard in their own right. NUREG-1805 lists the container becoming a projectile as the first hazard, before the fireball, and notes that pieces of the vessel may be thrown considerable distances. In quantitative risk assessment, "tank or equipment rupture or fragmentation" appears as a separate outcome under explosion overpressures.

For flammable contents, the fireball and its radiation are added to these hazards. The rapid failure of the container is followed by a fireball or major fire, which produces a powerful radiant heat flux. Fireballs from large accidents are usually of the order of 100 m in diameter and last only a few seconds, because the fuel mixes rapidly with the surrounding air. During that short time a fraction of the combustion energy is radiated, and the harm to people and property depends on both the intensity and the duration of the radiation. The maximum size is governed mainly by the mass of fuel released and vaporised and grows with the cube root of that mass; it does not depend on the storage pressure, provided the fuel vaporises. For propane at or above about 20 °C, complete vaporisation is expected.

A small test shows the time scale. When a cylinder holding 7.5 kg of LPG burst, the fireball reached a diameter of about 10 m within roughly 0.2 s. It stayed between about 10 and 12 m until the record ends at 1.8 s. The test went through three stages. First came rapid growth with a very bright flame, then a phase with little change in size in which the flame cooled from bright yellow to dull orange. In the final phase the fireball cooled rapidly and lifted off from the ground.

The fraction of the heat release that a fireball radiates is generally between 0.1 and 0.4. Large fireballs rise during their lifetime: for very large masses, above about 30,000 kg, even pressurised releases are dominated by buoyancy, and the fireball lifts off from the ground. The height of the fireball centre therefore matters when the distance to a target is worked out.

Estimating the radiation from a BLEVE fireball

NUREG-1805 names four parameters that are often used to judge a fireball's thermal radiation hazard: the mass of fuel involved and the fireball's diameter, duration and surface emissive power. For an object relatively far from the fireball, it gives a single empirical relation, after Hasegawa and Sato (1977) and Roberts (1982). In words: the radiant heat flux at the target grows with the mass of fuel vapour to the power 0.771. It falls with the square of the distance from the centre of the fireball.

$${\dot{q}}^{\prime }{\text{'}}_{\mathrm{r}}=\frac{828m_{\mathrm{F}}^{0.771}}{R^2}$$

Symbol	Description	Unit
${\dot{q}}^{\prime }{\text{'}}_{\mathrm{r}}$	thermal radiation from the fireball received at the target	kW/m²
$m_{\mathrm{F}}$	mass of fuel vapour in the fireball	kg
$R$	distance from the centre of the fireball to the target	m

The relation is dimensional: the constant 828 only gives the right result with the units in the table. The distance R is measured from the centre of the fireball, which lies above the ground; NUREG-1805 gives a separate method to find the fireball height from the volume of fuel vapour. The relation applies to targets "relatively distant" from the fireball. Close to a fire, point-source radiation methods overestimate the incident flux, because the near field depends strongly on flame size, shape and target orientation. NUREG-1805 presents them as suitable where a conservative estimate is acceptable.

Worked example. Take the LPG cylinder test above, with m_F = 7.5 kg. The numerator is 828 × 7.5^0.771 ≈ 3,915. At R = 20 m from the centre of the fireball:

$${\dot{q}}^{\prime }{\text{'}}_{\mathrm{r}}=\frac{828\times 7.5^{0.771}}{20^2}=\frac{3915}{400}\approx 9.8\mathrm{k}\mathrm{W}/{\mathrm{m}}^{\mathrm{2}}$$

At 30 m the flux is about 4.4 kW/m², and at 40 m about 2.5 kW/m². The fluxes are loads on the target, so they are rounded up. The same relation shows that the flux falls to 12.5 kW/m² at about 18 m, to 4.0 kW/m² at about 32 m and to 1.6 kW/m² at about 50 m (distances rounded up).

Radiant heat flux from the fireball of 7.5 kg of LPG against distance from the fireball centre, with damage thresholds. Based on Iqbal and Salley (2004), p. 5-15, Beyler (2002a), p. 3-307, and Barry (2002), p. 5-186.

These values are read against damage criteria from observations of large fires. In the table used by Barry, 12.5 kW/m² causes first-degree burns in 10 s and 1 % lethality in 1 minute. A flux of 4.0 kW/m² causes pain if the exposure lasts longer than 20 s, and 1.6 kW/m² causes no discomfort even for long exposure. The criteria are given for exposure times of 10 s or longer, while the fireball of the example lasted less than 2 s. The comparison therefore shows the importance of duration: the relation gives the intensity, but the harm also depends on how long the exposure lasts.

Incidents

Rail and road accidents with liquefied gases illustrate the scale of BLEVE fireballs. A report of the U.S. General Accounting Office cites a tractor-semitrailer carrying 34 m³ (9,000 gal) of LPG; about 2 minutes after the accident a fireball about 123 m in diameter was observed. Its radiant heat burned several people, a house, several other buildings and about 12 acres of woods.

In Oneonta, New York, a freight train of 27 cars derailed; 7 of them carried liquid propane, 120 m³ (33,000 gal) in total. Seconds after the derailment a large fireball erupted, believed to result from a tank car that split open. The fireball then heated other tank cars, which failed in several BLEVEs. The case shows how one fireball can heat neighbouring vessels until they fail in turn.

In Crescent City, Illinois, 15 cars of a freight train derailed, 10 of them each carrying 130 m³ (34,000 gal) of LPG. A tank car that rode over the pile tore a hole in another car, and the released propane produced the first fireball. Several later explosions followed over a period of hours and destroyed 24 dwellings and 18 businesses.

Use in risk assessment and protection

In quantitative risk assessment for the process industries, a "fire ball (BLEVE)" is one of the radiant heat outcomes considered. The others are the pool fire, the torch fire and the flash fire. The generic event trees of the World Bank, as presented by Barry, lead to a fireball when a release is instantaneous and ignited immediately. For pressurised storage and transport vessels, the typical failure case listed is "BLEVE (non-buried case only)", with a total rupture that is ignited. The results are drawn as effect zones around the release point. An example shows fatality and injury radii for a BLEVE fireball next to the zones for a torch fire, a flash fire and a vapour cloud explosion.

The mechanism shows where the weak spot of a vessel lies: the shell around the vapour space, which is not cooled by the liquid. For high-hazard locations where a fire develops quickly and reaches very high temperatures, for example with highly flammable fuels, NUREG-1805 notes that a deluge system is specified. In such a system all sprinklers or spray nozzles are open, and water flows from all of them at once when a detection system opens the deluge valve. The documents used here do not give water application rates for the cooling of vessels.

The BLEVE also appears in fire service procedures. A 1994 hazardous materials response plan of the Fire Department of the City of New York is reproduced in the NIST report on the World Trade Center response. It requires an evaluation before personnel are committed in potentially explosive situations. Among these it lists gas release or over-pressurisation of containers, with BLEVE given in brackets.

Limits of the simple approach

The fireball relation above is an empirical correlation and treats the fireball as a point source. It ignores the growth, rise and lift-off of the fireball, the angle between the target and the line of sight, and the absorption of radiation by the atmosphere. More detailed point-source and spherical models in the SFPE handbook include the radiative fraction, the atmospheric transmissivity and the geometry of the target. These belong to the topic of fireball modelling rather than to the BLEVE itself.

The blast and fragment effects are harder to quantify with the documents used here. Zalosh only refers to separate guidelines for BLEVE blast and fragment correlations. The failure temperature and fill level in NUREG-1805 are typical values quoted from the literature, not limits. They describe where most BLEVEs occur, not a temperature or level below which a vessel is safe. Finally, Barry gives the 8 to 15 minutes of exposure as a general characterisation of the BLEVE as a delayed event, not as a calculated failure time for a particular vessel.

See also

• Vapour cloud explosion
• Boilover
• Deflagration
• TNT equivalence

References

• Barry, T. F. (2002). Quantitative risk assessment in chemical process industries. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 5-176–5-213). National Fire Protection Association; Society of Fire Protection Engineers.
• Beyler, C. L. (2002a). Fire hazard calculations for large, open hydrocarbon fires. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-268–3-314). National Fire Protection Association; Society of Fire Protection Engineers.
• Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
• Lawson, J. R., & Vettori, R. L. (2005). The emergency response operations (NIST NCSTAR 1-8). National Institute of Standards and Technology.
• Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
• Zalosh, R. (2002). Explosion protection. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-402–3-421). National Fire Protection Association; Society of Fire Protection Engineers.