Hydrogen safety

Hydrogen safety concerns the assessment and control of fire and explosion hazards associated with hydrogen and its mixtures with an oxidiser. The central questions are whether a combustible mixture can form, whether it can ignite, and what consequences follow. The explosion-prevention principles described by Zalosh distinguish mixture composition from ignition and flame propagation. The EPRI/NRC hydrogen assessment applies that sequence to sources, release scenarios and exposed equipment in nuclear power facilities.

Hydrogen presents several features that affect this assessment. It is much lighter than air at ordinary temperatures, can ignite from weak sparks, and burns with a flame that may be difficult to see in daylight. Its flammable range is broad, but depends on the conditions under which propagation is tested. Those properties need to be connected to the particular release and enclosure rather than treated as a complete description of the hazard.

Source characterisation and release conditions

A hydrogen assessment begins with the source rather than a generic concentration limit. The EPRI/NRC report identifies quantity, concentration, temperature and pressure as conditions that vary between applications. Its nuclear-plant examples include storage, generator cooling, process piping and gas generation during battery charging. These applications do not create identical release histories or surrounding conditions.

Hydrogen source conditions and ignition timing define different scenarios. Based on Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-1–N-3 and N-7–N-11.

Stored hydrogen can be supplied as pressurised gas or cryogenic liquid. The report describes gas distribution from storage to users through piping. It separately discusses hydrogen generated within equipment or battery rooms. A finite stored inventory and continuing generation require different descriptions of the amount available during the scenario.

Potential release locations include valves, flanges, fittings and relief devices. The opening's position and direction can influence the subsequent flame or gas path. A relief discharge may have a more constrained direction than an unspecified equipment leak. The source description therefore needs both the released material and the physical arrangement through which it escapes.

The distinction between a release and its consequences also prevents premature assignment of a damage area. The report first characterises the hydrogen source, then develops scenarios and evaluates potential damage. Its simplified damage rules were devised for the nuclear facilities considered. They do not establish a universal separation distance for every hydrogen installation.

Flammable mixtures and propagation direction

The flammability limits describe the limiting fuel concentrations that permit flame propagation under specified conditions. The lower limit concerns a fuel-lean mixture, while the upper limit concerns a fuel-rich mixture with relatively little available oxygen. A measured limit therefore belongs to a fuel, oxidiser and test condition. It is not simply a property of the hydrogen inventory.

NUREG-1805 demonstrates the influence of propagation direction. Its room-temperature table gives a lower limit of approximately 4.1% hydrogen by volume for upward propagation, compared with 6.0% horizontally and 9.0% downwards. The listed upper value is 74% for all three directions. The surrounding discussion associates its 4.1% value with 25 °C, atmospheric pressure and saturated humidity.

Propagation direction in the cited room-temperature data	Lower hydrogen concentration	Upper hydrogen concentration
Upward	4.1% by volume	74% by volume
Horizontal	6.0% by volume	74% by volume
Downward	9.0% by volume	74% by volume

These differences show why a quoted concentration needs its test context. The higher downward-propagation limit does not make a mixture incapable of burning upwards. Nor does a room-average concentration identify every local mixture. The NUREG discussion of ventilation specifically recognises the possibility of trapped hydrogen pockets, particularly near a ceiling.

Dilution changes the mixture rather than removing the need for assessment. NUREG-1805 discusses added nitrogen, carbon dioxide and steam through flammability diagrams. The resulting combustible region depends on both fuel and oxidiser concentrations. A fuel concentration alone cannot describe the complete state when the oxidiser composition has also changed.

A mixture outside a flammable region can also change as it mixes with another stream. The handbook's flammability-diagram treatment follows composition through dilution or addition of gas. This makes the path between operating states relevant alongside the initial and final states. A single concentration measurement cannot describe every intermediate mixture involved in changing the atmosphere.

Ignition and the meaning of sensitivity

Minimum ignition energy refers to ignition under specified mixture and test conditions. NUREG-1805 shows that hydrogen's required spark energy varies with concentration and increases towards the flammability limits. Dilution can also change the energy requirement. A single quoted minimum therefore cannot describe every hydrogen-containing atmosphere.

Electrical sparks and small static discharges are among the ignition mechanisms identified in that report. Their significance follows from hydrogen-air mixtures being ignitable by weak sources under favourable conditions. This does not mean that every spark ignites every mixture. Composition, the energy delivered and the conditions at the ignition location remain part of the physical problem.

The EPRI/NRC report warns against inferring low ignition likelihood merely from a relatively high auto-ignition temperature. It discusses low-energy sources and the influence of materials present around the release. Spark ignition and heating to an auto-ignition condition are different mechanisms. A value associated with one does not exclude the other.

For lead-acid battery installations, NUREG-1805 identifies abnormal internal short circuits, heated terminals and leakage currents as possible ignition sources. Hydrogen generation and ignition can therefore be associated with the same equipment. That example supports examining equipment condition and local sources together. It does not provide a general electrical equipment classification scheme for every hydrogen application.

Immediate ignition, accumulation and explosion

Immediate ignition of a pressurised leak can establish a jet fire. The direction of the release then helps determine which targets receive direct flame exposure. The EPRI/NRC methodology treats this separately from storage-vessel failure and ignition of an accumulated cloud. The distinction preserves the different source geometries and consequence mechanisms.

Delayed ignition can involve a mixture that has travelled or accumulated before burning. A vapour cloud explosion scenario therefore requires examination of dispersion and the surroundings of the cloud. The report includes outdoor ignition and gas entering a building as separate cases. The eventual ignition location need not coincide with the original equipment leak.

Flame propagation also affects the pressure loading. Zalosh distinguishes subsonic deflagration from detonation propagating supersonically through the unburned mixture. Expansion and turbulence can increase the flame speed observed within an enclosure. A description based only on total combustible energy omits the rate and spatial development of that release.

The possibility of deflagration-to-detonation transition depends on conditions supporting flame acceleration. The handbook identifies mixture reactivity, elongated geometry, roughness and obstructions among the relevant influences. These factors require scenario-specific examination. The presence of hydrogen alone does not establish that a particular event will develop as a detonation.

Thermal and pressure effects also require different target descriptions. Jet impingement concerns local heating, while an explosion imposes a transient pressure history. The same item of equipment can be exposed differently in alternative scenarios. A consequence assessment must identify which mechanism its chosen model actually represents.

Ventilation, migration and connected spaces

Hydrogen's low density encourages upward movement, but the surrounding flow and geometry remain relevant. The EPRI/NRC report considers rising outdoor gas together with wind direction and atmospheric conditions. Inside buildings, it considers openings and ventilation connections that allow migration. Buoyancy is consequently one influence on the path, not proof that every release immediately leaves the affected area.

Ventilation connections can transfer hydrogen between release and ignition locations. Based on Electric Power Research Institute and U.S. Nuclear Regulatory Commission (2005), pp. N-8–N-11; Iqbal and Salley (2004), p. 16-11.

A ventilation intake can connect an outdoor release with an indoor hazard. The report describes examining concentration at the intake, mixing in the ventilation system and possible ignition positions. It separately considers gas moving between compartments through openings and ducts. A building boundary does not automatically interrupt the gas pathway.

Ventilation used to dilute hydrogen also needs a defined source rate and mixing assumption. NUREG-1805 discusses accumulation following loss of ventilation or conditions that permit local pockets. A flow expressed as air changes per hour does not by itself demonstrate uniform concentration. The distribution of supply and exhaust openings is part of the physical arrangement.

The report's battery-room discussion treats loss of ventilation as a condition requiring attention. Its nuclear-plant provisions and concentration values refer to the particular US documents and editions cited in 2004. Those historical values should not be combined into a universal current setpoint. The relevant operating requirement needs its own applicable document and installation basis.

An idealised concentration model

NUREG-1805 gives a concentration-buildup expression for a constant gas source in a ventilated enclosure. The following form uses its perfect-mixing case, with the mixing factor set to one. It starts with no hydrogen in the represented volume and assumes constant generation and ventilation. The result describes the modelled homogeneous concentration.

Ideal mixing gives an asymptotic concentration under constant conditions. Based on Iqbal and Salley (2004), p. 16-9, equation 16-3.

C = (G/Q)[1 − exp(−N)]

Symbol	Meaning in this idealised calculation	Unit
C	Hydrogen concentration as a volume fraction	dimensionless
G	Constant hydrogen generation or discharge rate	m³/min
Q	Constant ventilation airflow rate	m³/min
N	Number of theoretical air changes elapsed	dimensionless

Gas and air flow rates must use consistent volumetric reference conditions. Their ratio then gives a dimensionless concentration scale. Multiplication of C by 100 expresses the result as volume percent. Confusing a fraction with a percentage would change the reported value by a factor of one hundred.

As the number of air changes increases, the exponential term decreases and concentration approaches G/Q in this model. At the start, concentration is zero under the selected initial condition. These features explain the shape of the accumulation history. They do not establish the spatial concentration field in a real room.

For illustration, select hydrogen generation of 0.002 m³/min and airflow of 1 m³/min. The limiting model concentration is 0.002, equivalent to 0.20% by volume. After one, three and five theoretical air changes, the calculated values are approximately 0.126%, 0.190% and 0.199%. These selected inputs demonstrate the expression and are not ventilation design criteria.

Doubling the selected airflow to 2 m³/min halves the limiting concentration to 0.10%. At the same numbers of theoretical air changes, the corresponding values are approximately 0.0632%, 0.0950% and 0.0993%. This comparison holds generation constant and changes ventilation. It compares air-change counts, not identical elapsed times.

The expression contains ventilation flow in the denominator and cannot be evaluated by simply setting that flow to zero. Loss of ventilation requires a suitable accumulation balance and source history. Likewise, a changing source cannot automatically be represented by one constant value. Those changes require a different calculation or justified treatment of successive conditions.

Battery generation and limits of source models

Lead-acid batteries provide a source example in NUREG-1805. The report associates substantial hydrogen evolution with charging near full charge and describes the importance of overcharge conditions. Its generation estimate uses cell count, battery capacity and a float-current input. The current depends on the identified cell technology, age, charging voltage and temperature assumptions.

The report explicitly identifies the generation equation as a first-order approximation based on one vendor's recommendations. Its tabulated inputs concern stationary lead-acid cells. They are not evidence that every battery chemistry or operating condition releases the same amount of hydrogen. Applying the equation requires preserving the source's technological and operating scope.

Source uncertainty and room mixing uncertainty are separate issues. A carefully calculated room-average concentration can still use an unsuitable generation input. Conversely, an appropriate generation estimate cannot establish the absence of a local pocket if the mixing assumption is unsuitable. The source model and transport model require their own supporting evidence.

Isolation, exposure and interpretation

Extinguishing a visible flame and stopping hydrogen release are different outcomes. The EPRI/NRC discussion highlights the possibility of a flammable cloud forming if gas continues escaping without burning. Its treatment therefore connects isolation actions with the scenario sequence. A changed flame condition does not necessarily mean that the combustible inventory has been removed.

Hydrogen's weakly luminous flame also limits visual interpretation of an exposure. NUREG-1805 notes that the flame may be invisible in daylight. Meanwhile, the SFPE jet discussion treats radiative power and flame impingement as physical quantities requiring appropriate models. Visual brightness alone cannot provide those quantities or justify importing hydrocarbon-jet measurements unchanged.

Concentration monitoring provides information about the atmosphere at the measurement arrangement. NUREG-1805 identifies hydrogen measurement as part of implementing the battery-room protection measures it describes. The measured quantity remains distinct from a calculated room average or an assumed source rate. Interpreting one as the other would remove the spatial and operational assumptions behind the comparison.

For structural explosion effects, peak pressure is only part of the loading description. Zalosh also identifies pressure duration, reflection and impulse as relevant to dynamic response. A hydrogen scenario must therefore connect the chosen propagation model with the target's response. Reporting combustible mass alone leaves both that loading history and the response criterion unspecified.

The EPRI/NRC methodology explicitly separates scenario development from validation of computational models. Its appendix does not endorse one universal hydrogen consequence tool. A reported result therefore needs the source conditions, ignition branch, geometry, model assumptions and relevant validation. Those details define the consequence that has actually been assessed.

References

• Beyler, C. L. (2002). Fire hazard calculations for large, open hydrocarbon fires. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., Chapter 3-11). National Fire Protection Association.
• Electric Power Research Institute, & U.S. Nuclear Regulatory Commission. (2005). EPRI/NRC-RES fire PRA methodology for nuclear power facilities: Volume 2, Detailed methodology (EPRI 1011989; NUREG/CR-6850).
• Iqbal, N., & Salley, M. H. (2004). Fire dynamics tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission fire protection inspection program (NUREG-1805). U.S. Nuclear Regulatory Commission.
• Zalosh, R. (2002). Explosion protection. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-402–3-421). National Fire Protection Association.