Propagation

What propagation actually is

When a single cell goes into thermal runaway it self-heats through exothermic decomposition, vents flammable electrolyte gas, and releases a large pulse of heat — several megajoules for a modern 280-314 Ah LFP cell. Propagation is what happens next: that heat, plus burning ejecta and vent gas, drives adjacent cells above their own runaway onset temperature, so the failure cascades.

Engineers speak of propagation at distinct scales — cell-to-cell within a module, module-to-module within a rack, and rack-to-rack within a container — because the barriers, timescales, and test levels differ at each scale. Cell-to-cell steps typically take tens of seconds to a few minutes; an entire rack can become involved within the hour if nothing intervenes.

The dominant heat-transfer paths are conduction through busbars, cooling plates, and cell cans; radiation and hot-gas convection inside the module; and direct flame impingement once the vented gas ignites. Off-gassing is itself part of the propagation chain: the vent stream is hot, flammable, and can carry burning particulates into neighbouring modules.

The engineering goal is rarely to make a cell un-failable — across the millions of cells in a utility-scale plant, occasional single-cell failures are a statistical certainty even at very low defect rates. The goal is to break the transfer paths with inter-cell barriers such as aerogel or mica sheets, sealed module enclosures, and rack-level separation so a single-cell event does not become a full-enclosure event.

Why it matters in a grid-scale project

For a utility-scale asset the consequence of unchecked propagation is total loss of the enclosure, possible spread to adjacent containers, and — most dangerously — accumulation of explosive vent gas that can deflagrate when it finds an ignition source.

That last mechanism, not open fire, has caused the most serious BESS injuries to date, typically when responders opened or approached a gas-filled enclosure. The commercial consequences are equally real: insurability, permitting with the Authority Having Jurisdiction, inter-unit spacing on the site pad, and the Emergency Response Plan all hinge on demonstrated propagation behaviour.

Propagation resistance is also the core deliverable an integrator must prove to win a site. A design that contains a failure to a single module keeps the rest of the plant's POI capacity online and the project bankable; a design that propagates rack-to-rack can sterilize a large footprint and is effectively uninsurable.

This is why owners, lenders, and insurers increasingly require unit-level UL 9540A test data — a full-scale burn with instrumented adjacent target units — rather than cell-level certificates alone, and why a favourable unit-level result is worth real money in reduced separation distances, denser site layouts, and lower insurance premiums.

Typical values and standards

UL 9540A is the test method that quantifies fire and propagation behaviour at four levels — cell, module, unit, and installation — by forcing a single cell into runaway (typically with film heaters) and measuring heat release rate, heat flux to adjacent targets, vent-gas volume and composition, and whether runaway spreads beyond the initiating cell or module.

Its output is data, not a pass/fail certificate. UL 9540 is the separate system-level safety certification for the ESS product, and IEC 62619 covers cell and battery safety for industrial applications, with current editions adding their own thermal-propagation test. NFPA 855, the installation standard, is what converts UL 9540A data into separation distances, quantity limits, and protection requirements.

The default NFPA 855 envelope is instructive: roughly 0.9 m (3 ft) between ESS units and from walls, about 50 kWh per unit, and around 3 m (10 ft) from lot lines and exposures for outdoor installations — all modifiable based on large-scale fire test results.

On the gas side, a cell in runaway releases on the order of 1-2 litres of flammable gas per amp-hour of capacity depending on chemistry and state of charge, and LFP vent gas is hydrogen-rich. Hydrogen's Lower Explosive Limit is about 4% by volume, and NFPA 69 explosion-prevention systems are typically designed to hold the enclosure mixture below 25% of LEL; Deflagration venting to NFPA 68 covers the case where prevention fails.

Chemistry sets the baseline. LFP, the dominant stationary cell, typically has a higher thermal-runaway onset — often on the order of 200 °C and above, versus roughly 150-210 °C for NMC — and its cathode does not release oxygen, so it burns less energetically and is markedly more propagation-resistant. It is not propagation-proof: densely packed 280-314 Ah LFP cells store enough energy that cascade is entirely possible without engineered barriers, and their vent gas carries a higher hydrogen fraction, which shifts the dominant risk from fire toward explosion.

How it shows up in specs, studies and contracts

A working engineer meets propagation first in the UL 9540A test report package. Check which level was actually tested — cell-level data alone supports almost nothing at installation scale — and whether the tested configuration matches the product being quoted: cell model and capacity, barrier materials, module layout, and enclosure all matter, and a report for a 280 Ah design does not automatically read across to a 314 Ah successor.

Then read the unit-level outcome itself: no propagation beyond the initiating module, no flaming outside the unit, and acceptable heat flux at target walls are the results that unlock reduced spacing. Ask for the gas composition and volume data too, because the explosion-control design depends on it.

Propagation then reappears in the Hazard Mitigation Analysis submitted to the AHJ, in the fire-protection basis of design, in site-layout drawings — container spacing is often propagation-driven, not cable-driven — and in insurer questionnaires that ask explicitly for unit-level test outcomes.

Contractually, supply agreements increasingly warrant that delivered hardware matches the tested configuration, and a well-drafted contract requires notice of any cell or module change that would invalidate the test report. On the operations side, the Emergency Response Plan must reflect the tested behaviour, including how long responders should expect an enclosure to off-gas before any approach or ventilation decision.

Common pitfalls

The most common trip-wire is treating a UL 9540A report as a certificate. It is a data set; whether the design is acceptable is decided under NFPA 855 by the AHJ using that data; conflating the test method with the UL 9540 certification is a related error. A close second is configuration drift: integrators iterate cell suppliers and module layouts quickly, and a test report tied to last year's bill of materials may not cover what actually ships.

Third, propagation and explosion are coupled but distinct hazards — a design that never propagates thermally can still fill an enclosure with flammable gas from the single initiating cell, so gas management is mandatory even for genuinely non-propagating designs.