UL 9540A

What it is (precise)

UL 9540A defines four test levels: cell, module, unit (typically a representative rack, tested with instrumented target units), and installation. At the cell level, a single cell is driven into Thermal runaway, and the test measures the vent-gas volume, composition, lower flammable limit, burning velocity, and the temperature at which venting and runaway begin.

The module level then measures whether runaway cascades from cell to cell and the heat and gas release when it does; the unit level asks the same question between modules and racks; the installation level evaluates the fire-protection system at as-installed scale.

Through the 5th edition the protocol was sequential with an exit: each level carried performance criteria that determined whether the next level was needed, and because manufacturers designed systems to meet the unit-level criteria, installation-level testing was rarely performed in practice. The 6th edition reverses that stop logic — see the editions section below.

UL 9540A produces data, not a verdict. There is no "UL 9540A certified" status — UL itself states that the testing provides no certification, UL Mark, or pass/fail result. The product safety certification for the complete system is UL 9540, a separate standard covering the electrical, mechanical, and functional safety of the integrated ESS, and the fire codes require the ESS to be listed to that standard by a Nationally Recognized Testing Laboratory.

UL 9540A is the fire-characterization test whose report is consumed by the UL 9540 listing evaluation itself, by the Authority Having Jurisdiction, and by NFPA 855, the installation standard, to justify the as-built fire-protection design. Confusing the two standards is the single most common terminology error in BESS safety conversations.

Editions and the 2026 large-scale fire test

The method is binational (ANSI/CAN, US and Canada) and it moves quickly. The 4th edition, published November 12, 2019, carried the industry for five years; the 5th (ANSI/CAN/UL 9540A:2025) followed on March 12, 2025 and the 6th (ANSI/CAN/UL 9540A:2026) on March 13, 2026 — two editions twelve months apart. UL has stated that an effective date of January 1, 2027 will be established for the 6th edition, so 5th-edition reports legitimately remain in circulation through the transition. Every report cites the edition it was run to; read that line first.

The 6th edition's headline change sits at the top of the test stack. The sequence no longer stops when unit-level performance criteria are met: an installation-level large-scale fire test (LSFT) is now conducted regardless, unit-level testing is dropped for most lithium-ion systems — UL's guidance retains it for residential BESS and, per UL's service page, for non-residential systems with an active thermal-runaway propagation-prevention system — and the cell- and module-level tests remain mandatory.

The LSFT intentionally ignites the vented gases, assumes complete combustion of one enclosure, and places instrumented target enclosures at the manufacturer's minimum recommended spacing to detect runaway or hazardous temperatures. NFPA 855's 2026 edition newly requires large-scale fire testing and consumes exactly this data.

The driver, as the American Clean Power Association describes it, was field experience: fires occurred in products that had achieved favorable UL 9540A results, and root-cause data showed most failure events initiating outside the cells and modules the earlier editions focused on — so the LSFT treats the battery as a fuel source in a developed, worst-case fire rather than only as a potential initiator.

Two reading traps ride along with the new vocabulary. "Large-scale fire testing" has meant different things across NFPA 855 editions — the 2020 edition used the phrase for ordinary UL 9540A testing, the 2023 edition renamed that "fire and explosion testing," and from the 2026 edition LSFT means the whole-enclosure burn test — so date-stamp any LSFT claim before comparing sources.

In Canada, CSA published its own large-scale fire-test procedure as the technical specification TS-800:24, later incorporated into the consensus CSA/ANSI C800:25; UL, for its part, describes UL 9540A as the only consensus standard explicitly cited in NFPA 855 for large-scale fire testing, so treat the CSA route as an alternative an AHJ may accept rather than one the code names.

Why it matters in a real grid-scale project

The UL 9540A report is one of the gating documents for permitting a utility-scale site. NFPA 855 sets default separation distances between enclosures and from exposures such as lot lines, occupied buildings, and control rooms.

Large-scale test results demonstrating no unit-to-unit Propagation can be used to justify reduced spacing, back-to-back container rows, or higher energy density per acre — which directly affects how many MWh fit on a parcel and therefore project economics. On land-constrained sites, the difference between default spacing and test-justified spacing can be worth a double-digit percentage of installable capacity.

The report also drives explosion-protection engineering. Off-gassing from lithium-ion cells produces a flammable mixture, and the measured gas composition and generation rate are the design inputs that size NFPA 68 Deflagration venting panels or NFPA 69 explosion-prevention systems such as emergency exhaust ventilation designed to hold concentrations below 25 percent of the Lower Explosive Limit.

Insurers, lenders and offtakers commonly ask for the report before financial close, and carriers name UL 9540A in the minimum standard set they underwrite against, so a missing or unfavorable test can stall an otherwise interconnection-ready project.

Typical values and standards

Chemistry drives the results. LFP cells typically enter thermal runaway at higher onset temperatures (often in the range of roughly 200 to 270 degrees C at the cell surface, versus roughly 150 to 210 degrees C for NMC, though values vary widely with cell design and test method), release materially less heat per Ah, and show far less unit-to-unit fire propagation — a primary reason LFP dominates stationary storage.

NMC packs generally exhibit more energetic venting, sustained flame jetting, and higher propagation tendency, so their reports more often force larger separations, active suppression, or both. EV packs face the same physics but different standards, which is why automotive test data cannot substitute for a stationary UL 9540A report.

Vent-gas mixtures are dominated by hydrogen, carbon monoxide, carbon dioxide, methane, and heavier hydrocarbons plus vaporized electrolyte; hydrogen commonly makes up on the order of 20 to 50 percent of the flammable fraction and is usually the limiting species for explosion-protection sizing because of its low ignition energy and wide flammability range (roughly 4 to 75 percent in air).

Total vent volumes on the order of one to two liters per Ah of cell capacity are typical, but the project-specific report governs — vent volumes and propagation outcomes vary substantially with cell format, module packaging, and enclosure design.

The companion documents an engineer pairs with UL 9540A are: UL 9540 (system certification), UL 1973 (stationary batteries, cell through pack and rack), IEC 62619 (cell and battery safety for industrial applications, the common international analogue), NFPA 855 (installation; commonly cited defaults are roughly 3 ft / 0.9 m between ESS units and 10 ft / 3 m from exposures, both modifiable by large-scale test data), NFPA 68 and NFPA 69 (deflagration venting and explosion prevention), and the adopted fire code, typically the IFC. Each plays a distinct role; none substitutes for another.

How it shows up in specs, studies and contracts

In procurement, the UL 9540A report appears as a required vendor submittal alongside the UL 9540 listing, type-test certificates, and datasheets.

Check three things: that the tested cell model and revision match what will actually ship (cell suppliers change electrode designs between years), that the unit-level test used the same enclosure, rack layout, and fire-protection configuration as the offered product, and that the report is the full test report rather than a one-page summary letter — UL maintains a public UL 9540A database in which manufacturers can share as little as a model number with contact details or as much as the complete report, so a database entry confirms testing happened while only the full report shows what it found.

A report for a 280 Ah cell does not automatically cover the 314 Ah successor in the same container.

In permitting, the report is an input to the Hazard Mitigation Analysis that NFPA 855 requires when a design deviates from prescriptive defaults, and it underpins the Emergency Response Plan and fire-service consultations. The Authority Having Jurisdiction decides how much weight the data carries, and practice varies widely between jurisdictions; sophisticated AHJs in storage-heavy markets often ask pointed questions about test configuration.

In contracts, supply agreements increasingly include warranties that the delivered product matches the tested configuration, and insurers may condition premiums on specific report findings such as demonstrated non-propagation at unit level.

Common pitfalls

The most frequent trip-wires: treating "passed UL 9540A" in a vendor brochure as meaningful (there is no pass; ask what the data showed), accepting cell-level data as evidence of container-level safety (propagation behavior at unit level depends on the integration, not just the cell), and assuming a report transfers across product revisions or across integrators using the same cell.

Spec language deserves the same scrutiny: "UL 9540A certification" appears in vendor marketing and even in written requirements, and it names a document that does not exist — what a spec can legitimately require is the full test report, run to the applicable edition, with results acceptable to the AHJ.

Verify report currency too: with the 5th and 6th editions published twelve months apart in 2025 and 2026, a report that was current at financial close can be superseded before commissioning, and AHJs and insurers may reject reports run to superseded editions or to configurations that no longer match the shipped product. When in doubt, design to the specific report in hand, never to generic chemistry-level assumptions.