Why the second explosion does the damage

Facilities are rarely destroyed by the explosion that starts inside the equipment. They are destroyed by the one that starts everywhere else a fraction of a second later. Understanding that sequence changes both how these incidents are prevented and what an investigation has to establish.

The mechanism

A primary deflagration inside a collector, mill, dryer or duct produces a pressure wave that propagates into the surrounding space. That wave does two things simultaneously: it disturbs settled dust on horizontal and near-horizontal surfaces, entraining it into suspension, and it provides a flame front to ignite the resulting cloud. The suspended mass available in a poorly kept building can exceed the mass inside the process equipment by orders of magnitude.

The result is a spatially extended event rather than a contained one. Secondary explosions propagate through connected spaces, and in multi-storey facilities they characteristically travel upward through openings, lifting further accumulation as they go. This cascading behaviour is why damage patterns often appear disproportionate to the size of the initiating equipment.

How little accumulation is required

The intuition that a visibly clean facility is a safe one is unreliable. Guidance in NFPA 652 and its predecessor documents has long treated accumulation on the order of a small fraction of a millimetre over a meaningful proportion of floor area as sufficient to sustain a secondary event, depending on the material's bulk density and explosibility.

What this means practically is that surfaces which look merely dusty can hold enough material to matter, and that the relevant surfaces include many nobody inspects — the tops of ducts, beam flanges, light fixtures, cable trays, and the space above suspended ceilings. Investigations that consider only floor-level housekeeping frequently understate the available fuel.

Reading the damage

Because the secondary event is larger and later, it tends to overwrite the evidence of the primary one. This is the central difficulty of these investigations: the equipment where the incident began is often buried under, or destroyed by, the consequences of what it triggered.

Direction and sequence are nonetheless recoverable. Deformation of panels and structural members, the orientation of displaced equipment, differential scorching and the pattern of glazing loss all record the passage of pressure waves. Where deflagration venting operated, the state of the panels and the direction of discharge constrain the sequence considerably. Reconstructing this is painstaking, and it depends entirely on the scene being documented before debris is moved.

Isolation is the control that matters most

Explosion venting protects the vessel it is fitted to. It does not, on its own, prevent propagation into connected equipment or into the building. Isolation devices — chemical suppression barriers, rotary valves acting as flame arresters, back-blast dampers — are what interrupt the chain, and their presence, specification and functional state after the event are usually material questions.

A common finding is protection designed vessel by vessel without regard to interconnection. Ducting that ties several collection points together provides a propagation path that the individual vessel calculations never considered. Where that is the case, the design basis itself becomes part of the dispute.

What preservation looks like here

Because accumulation is the determinative variable, evidence of accumulation must be captured before cleanup. That means photographing surfaces at height, sampling residual layers in areas remote from the blast for depth and composition, and recording the condition of areas that were shielded from the event and therefore still reflect pre-incident housekeeping.

Those shielded areas are often the most valuable evidence available, precisely because they are the only remaining record of the condition the facility was actually in. They are also the first thing lost when restoration begins, and their loss is rarely deliberate.

The question behind the question

In litigation, secondary-event cases tend to resolve into whether the operator understood that housekeeping was a safety control rather than a matter of tidiness. Programmes that treat dust removal as a production or quality task, without a defined frequency tied to a hazard assessment, are difficult to defend once the mechanism is explained. Programmes that define accumulation thresholds, assign responsibility, and record execution are defensible even when an incident occurs.

Housekeeping as an engineered control

Where housekeeping is the primary defence against a secondary event, it has to be specified like any other safety control: a defined accumulation threshold, a defined inspection frequency derived from the rate at which the process actually generates fugitive dust, assigned responsibility, and a record of execution.

Programmes that fail tend to fail in the same way. Cleaning is scheduled by convenience rather than by generation rate, elevated surfaces are excluded because access is difficult, and compressed air is used to clear surfaces — which suspends the dust rather than removing it, creating precisely the condition the programme exists to prevent. Each of these is straightforward to establish after the fact and difficult to defend.

What modelling can and cannot settle

Computational modelling of dust dispersion and flame propagation has a place in reconstructing these events, particularly for testing whether a proposed sequence is physically plausible. It is most useful as a constraint on hypotheses rather than as a source of conclusions.

The limitation is input sensitivity. Results depend heavily on assumed accumulation depth and distribution, particle size, and the strength and location of the initiating event — none of which are known precisely after a destructive fire. A model calibrated against the physical evidence adds value; a model substituting for physical evidence tends not to withstand examination.

This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.