Fire testing

Fire testing is the controlled evaluation of materials, products, components or systems under specified fire or thermal exposure. Tests provide measurements such as ignition time, heat release rate, mass loss, flame spread, smoke production and fire resistance. Their purpose is to characterise response under defined conditions rather than to reproduce every possible real fire.

Tests range from bench-scale material measurements to full-scale product and system experiments. Each scale addresses different controlling processes, including material decomposition, room feedback, ventilation and installation geometry. Results should therefore be applied only with a suitable model and a clear account of the test arrangement.

Fire testing supports Reaction to fire and Fire resistance test assessments, design data, product comparison and model validation. It also reveals uncertainty from apparatus, specimen preparation, instrumentation and real-world variability.

Purposes and measurements

A test can establish whether and how a material ignites, how quickly it releases energy, how a flame spreads and how much smoke or toxic product is generated. Other tests evaluate whether a construction maintains integrity, insulation or structural adequacy. These are different performance questions and require different endpoints.

Heat release rate is central to many reaction-to-fire methods because it links material consumption with fire severity. Effective heat of combustion connects heat release and mass loss but can vary during a test. Bomb-calorimeter values alone therefore cannot represent all real burning behaviour.

Smoke and toxic-product measurements add further dimensions. Soot yield affects optical density, while carbon monoxide and other species influence inhalation hazard. Their yields vary with combustion mode, so test ventilation and exposure must be reported.

Tests can also provide model inputs. A heat-release curve, ignition criterion or material property can support calculations for a larger scenario, but the model still needs to represent scale, enclosure and installation. A measured result does not eliminate modelling assumptions.

Bench-scale calorimetry

The Cone calorimeter exposes a small specimen to controlled radiant heat and measures heat release rate by oxygen consumption. Its conical heater, spark ignitor, specimen holder and load cell sit beneath a hood and exhaust duct. Air and combustion products are extracted while oxygen and other gas concentrations are measured.

The standard specimen is 0.1 m by 0.1 m, with horizontal and vertical orientations possible. Maximum irradiance can exceed 100 kW/m², and the heater can be positioned above or beside the specimen. The initial exhaust flow can be adjusted between 10 and 32 litres per second, with relatively constant volumetric flow during a test.

Cone-calorimeter setup and oxygen-consumption measurement. Based on Janssens (2002), p. 3-55.

Oxygen-consumption calorimetry estimates heat release from oxygen used, commonly using 13.1 MJ per kilogram of oxygen consumed. Corrections can be needed for carbon monoxide and for substances whose oxygen-consumption constant differs. This is a measurement method, not a universal conversion from mass loss alone.

Bench-scale tests are efficient for comparing materials and obtaining input curves, but specimen size and boundary conditions differ from installed products. Orientation, backing, thickness and mounting can change heat feedback and decomposition. Application to a full object requires those differences to be evaluated.

Converting oxygen consumption to power

Oxygen-consumption calorimetry estimates released heat from the oxygen consumed by combustion. Drysdale gives an approximate conversion of 13.1 MJ per kilogram of oxygen for many common fuels. The basic relationship multiplies oxygen-consumption rate by that energy per unit mass. $$\dot{Q}\approx E_{{\mathrm{O}}_{\mathrm{2}}}{\dot{m}}_{{\mathrm{O}}_{\mathrm{2}},\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}}$$

Symbol	Description	Unit
$\dot{Q}$	Heat release rate	kW
$E_{{\mathrm{O}}_{\mathrm{2}}}$	Energy released per mass of oxygen consumed; approximately 13,100	kJ/kg
${\dot{m}}_{{\mathrm{O}}_{\mathrm{2}},\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}}$	Oxygen consumption rate	kg/s

For a 1 MW full-scale source, this simplified conversion gives 1,000/13,100 = 0.0763 kg/s of oxygen consumed. The calculation illustrates the principle at an item-fire scale; it is not a heat-release claim for a small cone-calorimeter specimen. The assumed conversion and the measured oxygen deficit must use compatible units.

An oxygen concentration alone does not give this mass-consumption rate. The instrument also needs the exhaust flow and the incoming-air reference, with corrections appropriate to the apparatus and gas measurements. Incomplete combustion, including carbon-monoxide production, can require corrections to the simple heat-per-oxygen relation.

Janssens describes the cone calorimeter with a 0.1 m × 0.1 m specimen under controlled irradiation. The measured heat release per specimen area can support comparisons between materials at that exposure. It does not by itself determine the burning area, ventilation or thermal feedback of a complete room fire.

Single Burning Item and room tests

The Single Burning Item test (SBI) uses two wall panels in a corner configuration without a ceiling and is a regulatory heat-release test for building products. It evaluates product response under that specified exposure. Results should not be interpreted as direct measurements of every wall or room installation.

A Room corner test places materials in a small enclosure with a defined burner and opening, then collects combustion products in a hood and exhaust duct. It includes hot-layer and room-surface feedback that open-burning tests omit. This makes it useful when enclosure effects are expected to influence heat release.

SBI and room-corner tests represent different boundary conditions. Based on Babrauskas (2002), p. 3-33 and Janssens (2002), p. 3-39.

Room tests can differ from open calorimetry as fire grows. Babrauskas reports limited augmentation for some upholstered chairs at lower heat release rates, but much stronger effects for mattresses and liquid pools. A single room correction is therefore not valid for every product.

Room-fire data also cannot be extrapolated directly to rooms with larger ventilation. Open-burning data may instead suit well-ventilated conditions when the fuel package is comparable. The source test and intended enclosure must be matched explicitly.

Large-scale and system tests

A Large-scale facade fire test evaluates facade systems at a scale where joints, cavities, cladding layers and external flames can interact. Such tests are relevant to system behaviour that cannot be inferred reliably from small material specimens. The actual method and installation remain necessary when applying the result.

Fire-resistance tests expose structural and separating assemblies to a defined furnace curve. Milke explains that ASTM E119 addresses structural integrity, temperature, flame and hot-gas passage and, in some cases, hose-stream response. These criteria are distinct from reaction-to-fire measurements of a material's own combustion.

Large tests can still be comparative rather than literal full-scale reproductions. Furnace dimensions, loading and end restraints may differ from the built structure. Actual performance requires appropriate analytical extension and validation rather than assuming the test building is identical.

C/VM2 (second edition, 2025) determines fire resistance ratings from standard tests or calculation methods verified by experimental data from standard tests. Its ratings report structural adequacy, integrity and insulation in that order. This is a regulatory use of fire-test evidence within New Zealand.

Data quality and uncertainty

Measurement uncertainty includes bias and random error. Calibration with a metered high-purity gas can reduce bias in oxygen-consumption systems, while interlaboratory trials help quantify random error. The 95% confidence intervals reported for some tests should be retained when comparing results.

Fire-test uncertainty comes from measurement and application. Based on Babrauskas (2002), pp. 3-32–3-33.

Apparatus features influence results. Heater spectrum and orientation, specimen mounting, exhaust flow and gas analysis can all affect the measured curve. Using different methods without checking those features can produce apparent material differences that are really method differences.

Specimen selection adds uncertainty. Composite products may have varying layers, moisture and manufacturing quality, and a small specimen may not represent the whole. Full-scale tests reduce some of these issues but introduce cost, ventilation and installation variability.

Data should include the complete time history and its definitions. Peak heat release rate, an early average and a long-term average answer different questions, and their values need not agree. The averaging interval and ignition reference must be stated.

Using results in engineering

Results can support material comparison, regulatory classification and design-fire selection. The intended use determines which measurements and acceptance criteria matter. A test that suits one purpose may not provide the data needed for another.

For reaction-to-fire assessment, heat release, ignition, spread and smoke describe how a product contributes to fire. For fire resistance, the construction's ability to continue separating or supporting loads is central. Confusing these measures can lead to incorrect claims about a material or assembly.

For modelling, bench-scale data can be extrapolated only with a defensible model. Full-scale data are stronger evidence when they match the actual fuel package and enclosure, but even then installation and ventilation differences need review. Similar product names are not enough.

For regulatory use, the applicable standard and edition must be identified. C/VM2 uses its 2025 second edition and New Zealand context, while ASTM E119 and ISO room tests have their own methods and criteria. A class or rating should not be transferred between systems without demonstrating equivalence.

Limits and common misunderstandings

A fire test does not prove performance in every real fire. It proves response under the reported exposure and arrangement. Extrapolation needs a model, compatible conditions and acknowledgment of uncertainty.

A larger test is not automatically a better test. Scale can introduce important room and system feedback, but it can also make results harder to reproduce and apply elsewhere. The appropriate scale depends on the decision and the controlling mechanisms.

A standard test is not a universal measure of noncombustibility or safety. It characterises particular properties, and other hazards such as toxicity, installation and changing ventilation may remain. All relevant performance criteria should be considered.

test data are not free of error merely because they follow a standard. Calibration, specimen preparation, instrument response and averaging still matter. These details belong with the data used for design or publication.

Test selection and review

Test selection should begin with the decision and its controlling mechanisms. Material-level ignition and heat release may be sufficient for a simple comparison, but an enclosure or assembly can introduce new feedback. The test should include those mechanisms without adding irrelevant complexity.

A test plan should specify the specimen, exposure, instrumentation and endpoint before testing. Mounting, backing, moisture, orientation and pilot location can all change the result. Recording these choices makes the resulting data reproducible and helps prevent an incorrect comparison with another method.

The exposure should be characterised as a time history where relevant. A constant radiant panel, a furnace curve and a developing compartment fire produce different boundary conditions. Using the first two as though they were interchangeable with the third would hide important transient behaviour.

Instrumentation should match the quantity required. Load cells measure mass loss, gas analysers measure species, exhaust measurements support oxygen-consumption calculations, and temperature and heat-flux instruments describe exposure. No one instrument supplies all of those outputs reliably without supporting measurements.

A test series can examine sensitivity rather than produce only one result. Changing heat flux, orientation, backing or ventilation reveals whether a conclusion remains stable. Those variations also identify the conditions under which extrapolation to the intended installation becomes uncertain.

Interlaboratory comparison adds another check. It can reveal apparatus, operator and analysis differences that a single-laboratory procedure would miss. Confidence intervals should be reported alongside the measured curve so that small differences between products are not overinterpreted.

Validation should compare like with like. A bench-scale heat-release curve should not be compared directly with a room-fire curve without accounting for enclosure feedback, and a layer-average temperature should not be compared directly with a near-ceiling reading. Definitions must match before agreement can be assessed.

A final report should separate measurement, calculation and acceptance. It states what was observed, which conversions were applied and which criterion was used to judge performance. This keeps the test result traceable when it is later used for classification, modelling or a design decision.

The Euroclass system illustrates why a test must be read through its regulatory method. A result becomes a classification only after the applicable procedures and criteria are applied, and another framework can use different evidence. The underlying measurement should not be assigned a class from an unrelated system without an equivalence assessment.

Acceptance criteria should be checked against the test's actual output. A furnace endpoint may be a temperature, a heat-release threshold, a spread observation or a retained structural function. Using the wrong endpoint can make a successful measurement answer an unsuccessful design question.

For product comparison, uncertainty should be included when differences are small. Calibration and specimen variation can create differences of the same order as the apparent product effect. Repeating the exposure or comparing independent laboratories can show whether the ranking persists.

For design, the complete evidence chain should be documented. It links the tested specimen and exposure to the conversion, model and acceptance criterion used in the building. Any weak link can change the conclusion even when the original measurement is sound.

A final review should also state which hazards were outside the test. Limited smoke, toxicity, ageing, installation damage and altered ventilation may not be represented. A fire-test result should therefore be used for the properties it measures, not as a guarantee of complete fire safety.

The test's role in the wider evidence chain should also be explicit. A result can support a classification, supply a model input or validate a predicted response, but those uses have different evidence requirements. A measurement that is adequate for one role may be incomplete for another.

See also

• Froude scaling
• Heat release rate
• Combustibility

References

• Australian Building Codes Board, National Research Council of Canada, International Code Council, & Department of Building and Housing New Zealand. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
• Babrauskas, V. (2002). Heat release rates. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Cadorin, J.-F., & Franssen, J.-M. (2003). A tool to design steel elements submitted to compartment fires: OZone V2. Part 1: Pre- and post-flashover compartment fire model. Fire Safety Journal, 38, 395–427.
• Drysdale, D. D. (2002). Thermochemistry. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Janssens, M. (2002). Calorimetry. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Kanury, A. M. (2002). Flaming ignition of solid fuels. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Kodur, V. K. R., & Harmathy, T. Z. (2002). Properties of building materials. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Lie, T. T. (2002). Fire temperature-time relations. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Milke, J. A. (2002). Analytical methods for determining fire resistance of steel members. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Ministry of Business, Innovation and Employment. (2025). C protection from fire: Verification method C/VM2 (2nd ed.). New Zealand Government.
• Mulholland, G. W. (2002). Smoke production and properties. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Purser, D. A. (2002). Toxicity assessment of combustion products. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Quintiere, J. G. (2002). Surface flame spread. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed.). National Fire Protection Association.
• Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.