Consequence Analysis

BLEVE Fireball Analysis

Understand the mechanisms of Boiling Liquid Expanding Vapor Explosions, model fireball dimensions and thermal radiation using CCPS and TNO correlations, and determine safe separation distances for personnel and occupied buildings.

API 2510

LPG Installations

Design standards for pressurized storage

CCPS

Consequence Analysis

Fireball correlations from test data

TNO Yellow Book

Physical Effects

Thermal radiation and dose modeling

Use this guide when:

  • Performing consequence analysis for QRA
  • Determining safe distances from LPG storage
  • Evaluating fire protection for pressurized vessels
  • Siting occupied buildings per API RP 752

1. BLEVE Mechanism

A Boiling Liquid Expanding Vapor Explosion (BLEVE) is the catastrophic failure of a vessel containing a pressurized liquid at a temperature significantly above its normal atmospheric boiling point. The term was coined by researchers at Factory Mutual in the late 1950s after investigating a series of devastating industrial incidents. While a BLEVE can occur with any superheated liquid, the most dangerous events involve flammable materials such as propane, butane, and LPG, which produce massive fireballs upon release.

Step 1

Fire Impingement

External fire heats the vessel wall. Liquid-wetted areas stay cool through boiling heat transfer, but vapor-wetted wall above the liquid level heats rapidly, losing strength.

Step 2

Wall Weakening

Steel loses approximately 50% of its yield strength at 600-700°C. The wall stress from internal pressure eventually exceeds the reduced wall strength.

Step 3

Catastrophic Rupture

The vessel tears open, typically in the vapor space. The sudden depressurization causes the superheated liquid to undergo rapid nucleation and flash to vapor.

Step 4

Fireball Formation

If the contents are flammable, the expanding vapor-liquid mixture ignites, forming a fireball that rises buoyantly while radiating intense thermal energy.

Critical distinction: A BLEVE is defined by the physical mechanism (rapid flashing of superheated liquid), not by combustion. Non-flammable liquids like water or CO2 can BLEVE, producing a blast wave and fragments but no fireball. However, the most severe industrial consequences occur when the contents are flammable, combining thermal radiation, blast, and fragment hazards.

Conditions Required for BLEVE

Three conditions must be present simultaneously for a BLEVE to occur. The liquid must be at a temperature above its atmospheric boiling point, meaning it is superheated relative to atmospheric pressure. The vessel must contain this liquid under pressure that prevents boiling. And the vessel must fail suddenly, allowing the liquid to depressurize and flash nearly instantaneously. The degree of superheat determines the violence of the event; higher superheat produces more rapid flashing and a larger fraction of the liquid converts to vapor.

The superheat limit theory (SLT) provides a thermodynamic threshold for BLEVE. If the liquid temperature exceeds the superheat limit temperature (approximately 89% of the critical temperature for many hydrocarbons), homogeneous nucleation occurs throughout the liquid bulk, producing an extremely rapid phase change that generates significant overpressure. Below the superheat limit, the flashing occurs more gradually through bubble nucleation at surfaces and impurities, producing a less violent event.

2. Fireball Correlations

Fireball dimensions are predicted using empirical correlations derived from experimental BLEVE tests ranging from small laboratory scale (a few kilograms) to large-scale tests involving several tonnes of propane and butane. The most widely used correlations relate the maximum fireball diameter and duration to the total mass of fuel involved.

CCPS / TNO Fireball Correlations: Maximum Fireball Diameter (meters): D_max = 5.8 × M0.333 Fireball Duration (seconds): For M ≤ 30,000 kg: t = 0.45 × M0.333 For M > 30,000 kg: t = 2.6 × M1/6 Where: M = Total mass of flammable material (kg) D_max = Maximum fireball diameter (m) t = Fireball duration (s) Fireball Center Height (buoyant liftoff): H_center = 0.75 × D_max (Roberts model) These correlations assume the entire fuel inventory participates in the fireball. In practice, 60-100% of the vessel contents typically contribute to the fireball.

Alternative Correlations

Several researchers have proposed different correlations based on varying datasets and assumptions. The Roberts (1981) correlation uses D = 5.8 * M^0.333, which is identical to the CCPS recommendation and is the most widely accepted. Moorhouse and Pritchard (1982) proposed D = 5.33 * M^0.327, which gives slightly smaller diameters. The Martinsen and Marx correlation, recommended for large inventories above 30 tonnes, adjusts the duration exponent to account for buoyancy effects that become dominant at large scales.

Mass (kg)Diameter (m)Diameter (ft)Duration (s)
10027882.1
1,000581904.5
5,000993257.7
10,0001254109.7
50,00021470215.5
100,00026988317.8

Fireball Dynamics

The fireball goes through three distinct phases. In the initial growth phase lasting approximately one-third of the total duration, the fireball expands to its maximum diameter near ground level. During the liftoff phase, buoyancy forces cause the hot combustion gases to rise, pulling the fireball upward and slightly reducing its diameter. In the final burnout phase, the remaining fuel is consumed as the fireball rises, and the luminous sphere gradually becomes an ascending cloud of hot combustion products. The maximum thermal radiation at ground level typically occurs during the early growth phase when the fireball is closest to the ground and at its largest diameter.

3. Thermal Radiation Modeling

Thermal radiation from a BLEVE fireball is modeled using the solid flame approach, which treats the fireball as a sphere with a uniform surface emissive power (SEP). The incident radiation flux at any receptor point depends on the SEP, the geometric view factor between the fireball and the receptor, and the atmospheric transmissivity along the radiation path.

Solid Flame Radiation Model: Incident radiation flux at receptor: q = SEP × F × τ Where: q = Incident radiation flux (kW/m²) SEP = Surface emissive power (kW/m²) F = Geometric view factor (dimensionless) τ = Atmospheric transmissivity (dimensionless) Surface Emissive Power: SEP = (f_rad × M × H_c) / (A_s × t) Where: f_rad = Fraction of energy radiated (0.25-0.40) M = Fuel mass (kg) H_c = Heat of combustion (kJ/kg) A_s = Fireball surface area = π × D² (m²) t = Fireball duration (s) Typical SEP values for hydrocarbon BLEVEs: 200-350 kW/m² (accepted industry range) View Factor for sphere: F = (R / L)² Where: R = Fireball radius (m) L = Distance from fireball center to receptor (m)

Atmospheric Transmissivity

Thermal radiation is absorbed by water vapor and carbon dioxide in the atmosphere as it travels from the fireball to the receptor. The Wayne (1991) model accounts for this absorption as a function of the path length and the partial pressure of water vapor. At short distances and low humidity, transmissivity approaches unity. At long distances or high humidity, transmissivity can drop to 0.5-0.7, significantly reducing the incident radiation. This effect is particularly important when calculating safe distances, as it provides some natural attenuation of the thermal hazard.

Radiation Fraction

The radiation fraction represents the proportion of total combustion energy that is emitted as thermal radiation rather than carried upward by convection. For hydrocarbon fireballs, experimental data suggests radiation fractions between 0.25 and 0.40, with propane and butane typically around 0.30. Larger fireballs tend to have lower radiation fractions because soot in the outer shell absorbs some outgoing radiation and re-radiates it inward. The radiation fraction is one of the most significant sources of uncertainty in BLEVE consequence modeling.

4. Safe Separation Distances

Safe distances from a potential BLEVE event are determined by the thermal radiation flux thresholds that correspond to specific injury or damage levels. These thresholds are used in facility siting studies per API RP 752 and quantitative risk assessments to determine minimum separation distances between LPG storage and occupied buildings, property boundaries, and emergency response positions.

Radiation (kW/m²)EffectApplication
1.6No harm with prolonged exposurePublic exposure limit, property boundary
4.7Pain threshold at 60 seconds exposureEmergency response perimeter
12.5First-degree burns, piloted ignition of woodMinimum building separation
25.0Second-degree burns, spontaneous ignitionEquipment damage zone
37.5Third-degree burns, significant structural damageLethal zone boundary
Duration matters: Unlike a steady-state fire, a BLEVE fireball is a transient event lasting only seconds. The thermal dose concept (radiation intensity raised to the 4/3 power multiplied by duration) provides a more accurate injury prediction than peak flux alone. Probit functions based on thermal dose are used in quantitative risk analysis to estimate probability of fatality as a function of distance.

Thermal Dose and Probit Analysis

For transient events like BLEVE fireballs, the probability of injury or fatality is calculated using the thermal dose and probit functions. The thermal dose combines the radiation intensity and exposure duration into a single parameter that correlates with injury severity. The Eisenberg probit model, widely used in QRA, relates the probability of fatality to the thermal dose through a probit equation. At a thermal dose of approximately 1000 (kW/m^2)^(4/3) * s, the probability of fatality for unprotected personnel is approximately 1%. At 2000, it rises to approximately 50%.

5. BLEVE Prevention

BLEVE prevention focuses on two strategies: preventing the conditions that lead to vessel failure, and mitigating the consequences if a BLEVE does occur. The primary prevention measure is adequate fire protection to keep the vessel wall cool during an external fire, preventing the temperature rise that leads to wall weakening and rupture.

Active Protection

Water Deluge

Fixed water spray systems providing 0.25 gpm/ft² minimum coverage per NFPA 15. Keeps vessel wall below critical temperature.

Active Protection

Depressuring

Emergency depressuring valves that reduce vessel pressure and lower the superheat, reducing BLEVE severity even if rupture occurs.

Passive Protection

Fireproofing

Cementitious or intumescent coatings that insulate the vessel wall, providing 1-2 hours of fire resistance per UL 1709.

Passive Protection

Pressure Relief

Properly sized relief valves prevent overpressure but cannot prevent BLEVE alone, since failure occurs below design pressure at elevated temperature.

Key Design Measures

  • Install fixed water spray (deluge) systems on all LPG vessels per NFPA 15 and API 2510
  • Provide adequate drainage to prevent pool fire formation under vessels
  • Install thermal relief valves sized for fire case per API 521
  • Consider mounded or buried storage to eliminate fire impingement exposure
  • Maintain adequate separation distances between vessels per API 2510 and NFPA 58
  • Implement fire detection and automatic deluge activation
  • Provide emergency depressuring capability for large vessels
  • Apply passive fire protection (PFP) to critical structural supports and vessel saddles

6. Fragment Hazards

In addition to the fireball and thermal radiation, a BLEVE produces vessel fragments that are projected at high velocity. Fragment throw distance is a significant hazard that often extends beyond the thermal radiation hazard zone. Historical data shows that BLEVE fragments from railway tank cars have been found up to 800 meters from the failure point, and in some events fragments have been projected over 1 kilometer.

Fragment characteristics depend on the vessel material, wall thickness, internal pressure at rupture, and the mode of failure. Ductile materials like carbon steel tend to produce fewer, larger fragments, while brittle materials produce more numerous smaller fragments. The number of fragments from a cylindrical vessel BLEVE typically ranges from 2 to 6 pieces, with end caps often being projected furthest due to the rocket effect from escaping pressurized contents.

Vessel VolumeTypical FragmentsMedian Throw (m)Maximum Throw (m)
< 5 m³2-4100-200300-500
5-50 m³2-5150-300500-800
50-200 m³3-6200-400700-1,200
Fragment analysis: Separate engineering methods are required to assess fragment hazards. The Baker-Strehlow-Tang method and TNO fragment models estimate throw distance, velocity, and impact energy. This analysis is complementary to the fireball thermal radiation analysis and should be performed as part of a complete BLEVE consequence assessment.

7. Historical Incidents and Lessons Learned

BLEVE incidents have provided critical data that shaped modern safety standards and prevention practices. The frequency of BLEVE events has decreased significantly since the 1970s and 1980s due to improved fire protection, better vessel inspection practices, and adoption of the standards discussed above. However, BLEVEs continue to occur when fire protection systems fail or are absent, reinforcing the importance of robust prevention and mitigation measures.

Key Lessons from Historical BLEVEs

  • Water cooling is essential: Virtually every major BLEVE incident involved either absent or inadequate water deluge coverage on the vessel that failed. Water spray at the design application rate can prevent BLEVE indefinitely.
  • Relief valves cannot prevent BLEVE: Pressure relief protects against overpressure but does not prevent the temperature-induced wall weakening that causes BLEVE. A vessel can BLEVE below its design pressure.
  • Vapor space exposure is critical: The unwetted wall above the liquid level is the vulnerable zone. As liquid boils off, the wetted area decreases and the vulnerable zone increases, creating a race between boil-off rate and fire exposure.
  • First responder positioning: Emergency responders must maintain separation distances based on the maximum credible BLEVE scenario. Approach should only be from the sides of a horizontal vessel, never from the ends (fragment throw direction).
  • Domino effects: A single vessel BLEVE can impinge on adjacent vessels, creating a cascading failure scenario. Separation distances and fire protection must account for these escalation scenarios.

Key Takeaways

Prevention

Water Deluge

Fixed water spray at 0.25 gpm/ft² is the single most effective BLEVE prevention measure.

Siting

Separation Distances

Use 4.7 kW/m² for emergency response and 1.6 kW/m² for occupied buildings.

Design

Multiple Layers

Combine active (deluge) and passive (PFP) fire protection with adequate relief and separation.