A Quiet Fuel Needs Loud Evidence: Detection Coverage at UK Hydrogen Refuelling Sites
Photo by Photo by Umair Dingmar on Unsplash on Unsplash
Hydrogen has an unusual safety characteristic: it is silent to the nose. Natural gas distributors add odorants so that a leak announces itself in a kitchen or on a street. That approach cannot simply be copied into fuel-cell hydrogen. Many of the sulphur compounds used to odorise natural gas are poisons to PEM fuel cells, and fuel-cell grade hydrogen specifications leave little room for added tracers. For UK refuelling sites, the consequence is straightforward: leak detection must be designed, maintained and documented as a primary safeguard rather than a decorative extra.
Odourisation is not a transferable fix
Operators sometimes ask whether a harmless odorant could be added to hydrogen as a simple warning measure. In practice, the answer is usually no. Even low-level contamination can affect stack performance, and the fuel supply chain is built around stringent purity limits. Detection must therefore sit outside the fuel path. That means sensors, sampling systems and alarm logic that do not introduce contaminants into the hydrogen stream, and a clear separation between instruments that monitor the atmosphere and equipment that touches the fuel.
This is not a reason to treat detection casually. It is a reason to treat it as an engineered system with defined coverage, response times and maintenance routines. As UK hydrogen refuelling expands in 2026, the sites that earn public and insurer confidence will be those that can show how their detection system works, not simply that one was installed.
Begin with a release map, not a sensor catalogue
Before choosing detectors, site managers should map where hydrogen can escape and where it can accumulate. At a typical refuelling site, the list may include compressor skids, storage manifolds, valve enclosures, dispenser cabinets, breakaway couplings, hose connections, vent stacks and tube-trailer interfaces. Each point has a different release geometry and a different maintenance access constraint.
Hydrogen is buoyant, but a high-pressure release can form a directional jet and can collect under canopies, inside cabinets or in ducting. A ceiling detector alone is rarely a coverage strategy. The release map should be informed by the site's DSEAR assessment, hazardous area classification and ventilation design, and it should be updated whenever equipment or layout changes.
Placement should reflect how hydrogen behaves
Detector placement is a compromise between risk, access and environmental exposure. Sensors should sit at likely leak origins and at likely accumulation points, with redundancy where a single detector failure would leave a blind spot. Enclosed compressor housings and dispenser cabinets need particular attention because small releases can concentrate quickly. Open canopies need a different approach: multiple detectors across the credible release zone, with attention to prevailing wind and thermal effects.
Complementary technologies can help. Acoustic or ultrasonic leak detection can identify the sound of a pressurised release before gas concentration rises, while point detectors confirm the atmosphere. Neither should be treated as a substitute for the other. The site should record why each technology was chosen, where it is installed and what it is expected to detect.
Alarm thresholds need a logic, not a default
Hydrogen's lower flammable limit is around 4% by volume. A two-stage alarm is common: an early warning at a fraction of the LFL and a higher alarm that initiates shutdown, isolation and ventilation. A 10% LFL pre-alarm and 20% LFL alarm may be encountered, but they are not a universal template. The operator should be able to justify the setpoints against the site risk assessment, the detector response time, the ventilation rate and the intended operational response.
The alarm matrix matters as much as the threshold. It should state who receives the alarm, what equipment shuts down, what must be manually reset, how the site is isolated and how emergency services are notified. A detector that triggers a display in an unmanned office is not a safety system. It is a data point. Detection coverage is only proven when the alarm produces a defined action within a defined time.
Maintenance is part of the safety case
Detectors drift, become poisoned and fail. Catalytic and electrochemical sensors can be affected by humidity, temperature, other gases and contaminants. Sampling lines can block. Filters can saturate. A detector that passes a calibration check but sits behind a blocked sinter is not coverage.
Maintenance routines should include functional checks, calibration against traceable gas, response-time checks and replacement schedules based on manufacturer guidance and site conditions. Records should capture the test gas used, the observed response, the technician, the date and any corrective action. Spare sensors and filters should be held for critical locations. The maintenance regime should also cover the alarm and shutdown pathway, not just the sensor head.
The evidence pack for safety cases and insurers
Whether a site sits within COMAH or a simpler DSEAR regime, the same evidentiary discipline applies. A useful evidence pack will typically include the current release map, detector layout drawings, coverage rationale, sensor datasheets, calibration certificates, alarm matrix, maintenance logs, drill records and change-control notes. It should also explain how the site prevents contamination of the fuel path and how it verifies that detection equipment has not become a contamination risk itself.
This pack should be written for someone who was not present at installation. An insurer, inspector or duty holder should be able to follow the logic from hazard to detector to alarm to response. If that chain is only in the installer's head, it is not evidence.
A practical checklist for site managers
- Confirm the release map is current and covers compressors, storage, dispensers, vents and trailer connections.
- Check that detector placement is justified by the DSEAR assessment and ventilation design, not by convenience.
- Verify that alarm thresholds are linked to a documented response plan and tested in drills.
- Review calibration certificates for traceability, response time and test gas.
- Inspect filters, sinters and sampling lines for blockage or contamination.
- Confirm that detection equipment cannot introduce contaminants into the fuel stream.
- Keep change-control records for layout, equipment and alarm-logic modifications.
- Hold spare sensors and critical components for high-risk locations.
Hydrogen's lack of smell is not a detail to be managed with signage. It is a design constraint. UK refuelling sites need detection systems that are placed with intent, maintained with discipline and documented with enough clarity to withstand scrutiny. The fuel may be quiet, but the evidence around it should not be.