Explosion and Pressure-Vessel Failure

A BLEVE requires two conditions: a liquid stored above its atmospheric boiling point (meaning it is superheated by the pressure of the vessel), and a sudden loss of containment. The most common trigger is external fire heating the vessel wall above the metal's yield strength at the liquid-vapour interface, where liquid cooling is absent and the wall loses structural integrity. When the wall tears, the pressure drops instantly to atmospheric. The superheated liquid flashes to vapour, expanding rapidly and producing the characteristic pressure wave.

If the liquid is flammable (LPG, liquefied natural gas, fuel oil), the released vapour cloud ignites and produces a fireball whose diameter and duration are roughly proportional to the mass of fuel involved. The fireball is a thermal hazard distinct from the overpressure wave, and the two together produce the combined blast-and-thermal effects associated with large industrial BLEVEs such as the Mexico City PEMEX LPG facility fire of 1984.

• Fragment evidence: vessel fragments are propelled as missiles, often landing hundreds of metres from the failure point. Fragment landing positions, mapped by survey, allow back-calculation of minimum internal pressure at failure using ballistic models.
• Fracture surface: the tear in the vessel wall begins at a stress-concentration point, often a weld seam or a corrosion pit. Fractographic analysis of the fracture surface can reveal whether failure was ductile (material had remaining plastic capacity), brittle (rapid fracture at stress below yield), or caused by a pre-existing fatigue crack.
• Overpressure damage field: the pattern of structural damage to surrounding buildings, particularly inward collapse of blast-facing walls, allows estimation of peak overpressure at various distances using empirical blast-scaled distance correlations.

Explosions in industrial and forensic contexts are almost always either deflagrations or detonations, and the distinction matters both for understanding what happened and for interpreting the structural damage. A fuel-gas leak that accumulates in a building and ignites typically deflagrates. A high explosive placed as a charge detonates. The two produce fundamentally different pressure-time histories and different damage patterns.

In practice, a deflagration can transition to detonation (DDT) in a long pipe or duct if the conditions are right, producing a hybrid damage sequence that can confuse the initial assessment. Investigators use the spatial pattern of window failure, wall direction of displacement, and fragment velocities to reconstruct the pressure-time history and determine whether DDT occurred. Soot deposition patterns and burn marks on surfaces also help distinguish the thermal phase of a deflagration from the blast-only effects of a detonation.

Not all pressure-vessel explosions are caused by single acute events. Many result from long-running degradation mechanisms that reduce the vessel's effective wall thickness or introduce crack initiation sites, until the vessel can no longer sustain normal operating pressure. The two most common mechanisms in industrial process plant are cyclic fatigue and corrosion under insulation.

Cyclic fatigue in pressure vessels operates on the same principles as in any other structure. Vessels that undergo repeated pressurisation and depressurisation cycles (start-up and shutdown cycles, process pressure fluctuations, hydraulic surge) accumulate fatigue damage at stress-concentration points such as nozzle attachments, weld toes, and thickness changes. ASME Boiler and Pressure Vessel Code Section VIII Div. 2 requires fatigue analysis for vessels above specific cycle counts, but many older vessels were designed to earlier codes without explicit fatigue provisions.

Corrosion under insulation (CUI) is a concealed external corrosion mechanism. Water from rain, process leaks, or condensation penetrates the thermal insulation, becomes trapped against the metal surface, and drives corrosion at rates that depend on metal temperature, water chemistry, and oxygen availability. The most aggressive CUI zone for carbon steel is roughly 60-120 degrees Celsius, where the surface stays wet but is warm enough to accelerate the electrochemical reaction. Stainless steel suffers stress-corrosion cracking under CUI rather than simple pitting.

The ASME Boiler and Pressure Vessel Code (BPVC) sets design rules for the construction of pressure vessels in the United States and many international jurisdictions. It is organised into sections by vessel type: Section I for power boilers, Section VIII for unfired pressure vessels (Divisions 1, 2, and 3), and Section X for fibre-reinforced plastic vessels, among others. The code specifies design margins, material allowable stresses, weld joint efficiencies, and required inspection and testing at fabrication.

API 579 (Fitness-For-Service) operates at the other end of the vessel life cycle. It provides assessment methods for a vessel that has developed a defect, corrosion, or damage during service, allowing engineers to determine quantitatively whether continued operation is safe and under what conditions. Relevant sections include Level 1, 2, and 3 assessments for general metal loss (corrosion thinning), pitting, and crack-like flaws. The standard is widely used both as an engineering tool and as a reference for forensic assessment of why a vessel that had been inspected and cleared subsequently failed.

On 23 March 2005, workers at BP's Texas City refinery in Texas were restarting the isomerisation unit after a maintenance shutdown. The raffinate splitter tower was being filled with hydrocarbon feedstock and brought to operating temperature. During the startup, liquid levels in the tower were not correctly monitored. The tower was overfilled, and hot liquid hydrocarbon was routed through the overhead system into a blowdown drum that was already full.

The blowdown drum, a cylindrical vessel designed to receive vapour and small amounts of liquid from process upsets and vent them to atmosphere through a stack, had no liquid-level instrumentation. Liquid overflowed the stack and fell to the ground as a flammable pool. Vapour from the liquid formed a cloud that found an ignition source, likely a running vehicle engine nearby. The deflagration ignited the liquid pool and the resulting explosion and fire killed 15 people in the trailer complex adjacent to the unit and injured 180 others.

The US Chemical Safety Board (CSB) investigation, published in March 2007, identified multiple layers of failure: the blowdown drum design with no liquid-level indication, the open-to-atmosphere stack that discharged flammable vapour at grade level rather than to a flare system, a startup procedure that was not followed, a level-indicator on the tower that was defective, and organisational factors including cost-cutting and inadequate safety culture. The incident drove changes to ASME process-safety design guidance and API standards for blowdown systems, including requirements for closed systems (flare or scrubber connection) rather than open atmospheric stacks in hydrocarbon service.

An explosion investigation proceeds through a defined evidence-collection and analysis sequence that mirrors the broader failure-investigation process described in ASTM E860.

1. Scene documentation and preservation
   Total-station and photogrammetric survey of fragment positions relative to the vessel origin. GPS coordinates and photographs for each fragment. Documentation of structural damage to surrounding buildings, including wall displacement directions, window failure patterns, and damage gradients with distance.
2. Fragment recovery and cataloguing
   Each fragment is recovered, weighed, and its original position on the vessel tentatively identified from wall-thickness marks, weld seams, and nozzle locations. Reconstruction of the vessel from fragments, even partial, establishes the geometry of the failure initiation zone.
3. Fracture surface examination
   The fracture surface at the initiation zone is examined macroscopically for beach marks, fatigue striations, corrosion pits, or hydrogen embrittlement features. Scanning electron microscopy provides micrometre-scale detail. Material composition is verified against the original mill certificate.
4. Pressure reconstruction
   Fragment ballistic analysis and overpressure damage assessment at known distances are used to estimate minimum internal pressure at failure. The estimated pressure is compared against the vessel design pressure and the operating pressure at the time of failure from process records.
5. Failure mode determination
   The physical evidence, the pressure reconstruction, and the operational history are synthesised to establish whether the cause was material deficiency, design inadequacy, operating exceedance, external fire (BLEVE), or latent corrosion or fatigue. The determination is tested against the ASME BPVC requirements applicable to that vessel.