- 13 Jan 2026
- 8 Minutes to read
Hydrogen indoor release and explosion
- Updated on 13 Jan 2026
- 8 Minutes to read
Hydrogen refuelling stations (HRS) are a critical element of the energy transition, enabling the deployment of hydrogen-powered vehicles for public and commercial transport. To achieve sufficient onboard vehicle range, hydrogen is typically compressed to very high pressures (700–900 bar) prior to storage and dispensing. As a result, HRSs include high-pressure hydrogen compressors or pumps, often housed in enclosed or semi-enclosed modules for noise control, weather protection, and operational reasons.
While an enclosure offers practical advantages, it significantly increases the potential consequences of a hydrogen leak. Hydrogen has a wide flammability range in air (4–75% by volume) and exhibits its highest reactivity at concentrations between approximately 25–50% by volume, where laminar burning velocities are maximal. In confined and congested environments, the accumulation of hydrogen within this concentration range can result in severe explosion overpressures if ignition occurs.
Scenario
Hydrogen leak inside the compressor enclosure |
This scenario considers a continuous hydrogen release originating from the discharge side of a high-pressure compressor (e.g. from flanges, valves, connections) located inside an enclosure. The leak occurs during the compressor's operation. Mechanical ventilation is assumed to be operational at the time of the release, located at the top of the container. ![]() Location of leak on discharge of compressor [1] |
The objectives of this scenario are to:
Assess hydrogen dispersion and accumulation inside the enclosure.
Assess the size of the potential flammable gas cloud.
Evaluate the influence of the mechanical ventilation, considering a 1000 m3/h ventilation rate.
Calculate the corresponding explosion overpressure in case of ignition.
Assess flammable concentration outside of the enclosure.
Modelling approach
To simulate the case, we start at the top of the tree and continue downward, filling in information node by node:
Add background. You can use tile servers, pictures or drawings.
Add equipment. Set the location of the compressor.
Select models. Start from the point of release and add subsequent models in the order of events to create a model tree.
Release model
To determine the outflow conditions at the exit (e.g. flowrate, temperature, pressure) we apply the Gas release model. These outputs will be used as input for the subsequent Indoor Ventilation, Dispersion and Overpressure model. Leak size is selected based on hazardous area classification guidance and falls within the range suggested by BS EN 60079-10-1. Leak definition:
Leak area: 0.5 mm²
Leak duration: 30 s (until isolation)
Leak direction: Vertical downwards (worst case) and horizontal
Leak location: Compressor discharge piping
The vertical downward orientation is considered conservative, as jet momentum is rapidly dissipated upon impact with the floor, promoting lateral spreading and accumulation before buoyant rise. All input parameters are detailed in the table below.
Process Conditions | Leak 30s, 0.8 mm |
Chemical name | HYDROGEN (DIPPR) |
Initial temperature in vessel (°C) | 30 |
Calculation Method |
|
Use which representative rate | First 20% average (flammable) |
|---|---|
Type of vessel outflow | Release through hole in vessel |
Type of release duration | Calculate until specified time |
Maximum release duration (s) | 30 |
Expansion type | Isothermal |
Process Dimensions | |
Vessel volume (m3) | 0.25 |
Hole diameter (mm) | 0.8 |
Hole rounding | User defined |
Discharge coefficient (-) | 0.85 |
Indoor modelling
In the next step, the Indoor Ventilation, Dispersion and Overpressure model can be applied to predict gas accumulation and possible overpressure from the ignition of a flammable cloud in an enclosure. The mechanical ventilation system is defined as follows:
Location: Top of enclosure
Extraction rate: 1000 m3/h → Refreshment rate 62/h (based on room dimensions)
Area: 1.21 m2 (1.1 × 1.1 m)
Release orientation and turbulent jet interaction with the wall processed by Indoor Ventilation, Dispersion and Overpressure model | |
|---|---|
![]() Vertical downward release | ![]() Horizontal release |
Modelling choices:
Overpressure calculation: Selecting “Yes” for ‘‘Use overpressure model’’ activates overpressure calculation and introduces an additional input parameter: Adiabatic flame temperature. This value is not available in the EFFECTS chemical database but can be obtained from external engineering references. In this case, an adiabatic flame temperature of 2483 K was used, based on data from The Engineering ToolBox.
Scenario duration: The total simulation duration for indoor calculation was extended to 100 s in order to monitor the behaviour of the gas after the leak isolation, while the leak itself was assumed to last 30 s. This allows the dispersion and dilution of hydrogen in the enclosure to be assessed beyond the release phase.
Type of release: Due to the strongly buoyant behaviour of hydrogen, a jet release was selected as a type of release. This activates the two-layer model, which predicts the formation of a hydrogen-rich layer near the ceiling of the room, consistent with expected hydrogen accumulation behaviour in confined indoor spaces. The jet release type can also be applied to heavy gases with negative buoyant behaviour, where an inverted two-layer model is applied.
Model inputs:
Calculation Method | Downward release | Horizontal release |
Use overpressure model | Yes | Yes |
|---|---|---|
Adiabatic flame temperature (K) | 2483 [2] | 2483 [2] |
Ventilation exhaust representative rate | First 20% average (flammable) | First 20% average (flammable) |
Type of release | Jet release | Jet release |
Source Definition | ||
Duration of the scenario (s) | 100 | 100 |
Process Dimensions | ||
Hole diameter (mm) | 0.8 | 0.8 |
Hole rounding | User defined | User defined |
Discharge coefficient (-) | 0.85 | 0.85 |
Height of release (Z-coordinate) (m) | 1 | 1 |
Outflow angle in XZ plane (0°=horizontal; 90°=vertical) (deg) | -90 | 0 |
Distance to wall (m) | 1 (distance to ground) | 2.4 (farthest distance to side wall) |
Width of the room (m) | 2.4 | 2.4 |
Length of the room (m) | 2.5 | 2.5 |
Height of the room (m) | 2.7 | 2.7 |
Room temperature (°C) | 30 | 30 |
Ventilation type | Mechanical | Mechanical |
Temperature of the ventilation (°C) | 30 | 30 |
Exhaust area (m2) | 1.21 | 1.21 |
Exhaust height from floor (m) | 2.7 | 2.7 |
Refresh rate (/hour) | 62 | 62 |
LIMITATIONS:
There may be scenarios where local extractors inside the container or explosion relief panels are installed. These features cannot be explicitly modelled in EFFECTS. The indoor dispersion model in EFFECTS assumes an empty room with ventilation openings located on the side walls. To accurately represent complex explosion mitigation and ventilation measures such as explosion venting panels, detailed ventilation flow patterns, and localized extraction systems, advanced CFD modeling can be applied using FLACS.
Outdoor dispersion
In some cases, it may also be necessary to assess whether a flammable cloud can form outside the enclosure and subsequently encounter an external ignition source. To evaluate dispersion beyond the enclosure, the Dispersion model can be used.
The dispersion model allows the definition of a user-defined window, which represents the ventilation or other opening through which gas exits the enclosure. The model automatically processes outputs from the Indoor Ventilation, Dispersion and Overpressure model, and transfers the representative outflow rate and chemical mass fraction in the released stream to the dispersion calculation. Input parameters for the model are detailed in a table below.
Downward release | |
Type of release | User-defined window |
Source Definition | |
Mass flow rate of the source (kg/s) | 0.015517 (Linked from Indoor Ventilation, Dispersion and Overpressure model) |
Building wake | Yes |
Width source in crosswind (y) direction (m) | 1.1 (Relates to the effective width of the exhaust) |
Height source in vertical (z) direction (m) | 0.1 (Relates to the effective height of the exhaust, adjusted for vertical release) |
Representative mass fraction chemical (-) | 0.013 (Linked from Indoor Ventilation, Dispersion and Overpressure model, adjusts the concentration of the chemical in the source mass flow rate) |
Process Dimensions | |
Offset X direction (distance) start dispersion (m) | 0 *(Depends on the position of the vent to the leak location) |
Height of release (Z-coordinate) (m) | 2.7 (Height of the vent) |
Vent position
The User-defined window models horizontal releases only. However, in many practical situations, gas may exit a structure through a roof vent or chimney, resulting in a vertical release.
If a vertical Jet release cannot be applied in the dispersion model due to low momentum, the User-defined window may be used as an approximation. In such cases, the user must be aware that the release will be treated as horizontally oriented, which can significantly affect the dispersion pattern and predicted concentrations.
.png)
Illustration of the connection between the indoor and outdoor dispersion models in EFFECTS.
Results
Indoor Ventilation, Dispersion and Overpressure model reports 3 types of results:
Representative values - characteristic conditions at the outflow, providing a simplified description of the state of the system, representing the full release duration.
Ventilation - values at the exit from the room.
Room or layer properties - describe the characteristics of the stratified gas layer formed during the simulation or the state of the room.
In graphs of Indoor Ventilation, Dispersion and Overpressure model for vertical downward release below, we can see that a flammable hydrogen cloud in the room is formed within seconds of leak initiation, and the room ventilates approximately 25 s after the isolation of the leak takes place.
Indoor Ventilation, Dispersion and Overpressure model: Flammable mass in a cloud; Vertical downward release | Indoor Ventilation, Dispersion and Overpressure model: Layer thickness; Vertical downward release |
|---|---|
![]() Indoor Ventilation, Dispersion and Overpressure model: Flammable mass of cloud; Vertical downward release | ![]() Indoor Ventilation, Dispersion and Overpressure model: Layer Volume; Vertical downward release |
The expected overpressure is displayed as a function of time, depending on the time of ignition, reaching its maximum at 30 s overpressure of 6 mbar.
.png)
Indoor Ventilation, Dispersion and Overpressure model: Overpressure at ignition time; Vertical downward release
The values in the report of the model can be used to communicate key parameters to subsequent simulations. However, to properly analyse the dynamic behaviour of the release and the dispersion process in the room, it is recommended to examine the time-dependent graphs, which provide a more complete representation of the transient evolution of the gas cloud.
If the cloud does not ignite and is released to the outdoor environment, it is expected to reach a distance of approximately 11.2 m at 25% of the lower flammability limit (LFL) and 5 m at 50% LFL. This indicates the potential for a flammable cloud to form outside the enclosure, presenting a possible outdoor hazard if an ignition source is present.
.png)
Dispersion - Flammable cloud model: Side plume; Vertical downward release
Conclusion
This use case demonstrates how EFFECTS can be applied to assess dispersion and ignition of flammable clouds indoors. By combining the Gas release, Indoor Ventilation, Dispersion and Overpressure model and outdoor Dispersion model EFFECTS enables a coherent evaluation of hydrogen accumulation, flammable cloud formation, and potential explosion loads within confined spaces and in their surrounding environments.
The results show that a flammable hydrogen cloud can form within seconds of leak initiation inside the enclosure and that dilution is primarily achieved after leak isolation, despite mechanical ventilation. The modelling further indicates that hydrogen may be released to the outdoor environment, where flammable concentrations can extend several metres from the enclosure, representing a potential external ignition hazard.
While EFFECTS does not explicitly model detailed internal features, the tool provides a robust and transparent methodology for screening-level consequence assessment and for transferring representative outflow conditions to outdoor dispersion analyses. Overall, the study highlights the importance of rapid detection and isolation in HRS compressor enclosures and demonstrates the capability of EFFECTS to support safety evaluations for hydrogen infrastructure in the energy transition.
Download the project file
Explore the project file simulating high-pressure hydrogen release in the refuelling station (HRS) compressor enclosure.
To view the project file, please open it using the EFFECTS software. If you don’t have the software, you can download and use the free viewing demo version of EFFECTS via the link below.
References:
1. Connor, P., et al. (2022). Analysis of Flammable Cloud Formation in a Hydrogen Compressor Shelter. Presented at IChemE Hazards 32 Conference.
2. The Engineering ToolBox (2005). Adiabatic Flame Temperatures. [online] Available at: https://www.engineeringtoolbox.com/adiabatic-flame-temperature-d_996.html [11-12-2025].
.png)
.png)
.png)
.png)
.png)