Hydrogen relase from a pressure regulator

A hydrogen release occurred on a pressure regulator installed on a standard 0.5 kg hydrogen cylinder.
When solenoid valve was opened to allow hydrogen to flow, a loud noise was heard and hydrogen began to flow out of the pressure relief valve installed on the side of the regulator.
The low-pressure gauge on the regulator was indicating its max value (>200 psi, > 14 bar).

The valve on the bottle was shut off, stopping the flow. The released hydrogen did not ignite.

Cause of the failure was an elastomeric O-ring sealing the internal "nozzle" to the "seat assembly." It was deformed and had extended into the nozzle orifice, preventing the seat assembly from properly seating. This allowed high pressure hydrogen to continuously flow into the low-pressure side of the regulator. Nothing downstream of the regulator was damaged.

The INITIATING CAUSE was a damaged O-ring and the consequent incapacity of the regulator to guarantee pressure control and reduction
It is not known what led to the failure of the elastomer ring.
According to manufacturer of the pressure regulator, it happened very rarely that the elastomeric ring was damaged. The ring material was specifically selected for hydrogen/methane service and subjected to tens of thousands of open-close cycles.

The regulator has a pressure relief valve as protection and it operated properly, relieving the pressure in the system. In this sense, the mitigating measure in place worked as planned, and this event can be classified as small incident, borer line to a near miss.
The manufacturer did not require maintenance of these pressure regulators.
The ROOT CAUSE is to attribute to small shortcoming of the design of the hydrogen supply installation, because a pressure relief device should not discharge at the level of the cylinders, where the presence of personnel is unavoidable.

O-ring, PRD

The pressure regulator was of the single-stage type, and had functioned properly prior to the event through several on-off cycles. The pressure was in the range 150 to 200 bar.

In this event, the failure of the pressure regulator to guarantee effective pressure reduction caused the system downstream of it to be exposed to a pressure higher that the maximal design supply. Already before opening the solenoid valve which initiated the hydrogen flow, the pressure gauge at the low-pressure side of the regulator was indicating a max reading of above >200 psi (14 bar). Probably, up to the solenoid valve, the pressure on the low-pressure of the regulator side was already at a too high value but had not yet reached the set value of the pressure relief device.

This safety measure worked as designed, avoiding escalation of this small incident. Moreover, the small size of the orifice in the regulator was delivering a very low mass flow rate. Both aspects provided an adequate protection to the rest of the system.

Nevertheless, as highlighted byH2TOOLS, this event provided several lessons:
1. A pressure regulator is not a safety device. If failing in regulating the pressure, downstream components are exposed to the pressure present in the bottles or cylinders. If these components are not rated for that pressure value, they will be damaged. Therefore, it is recommended that pressure safeguards protection be added to the system.
2. The fact that the hydrogen cylinder was placed outside facilitated hydrogen dispersion, in an open space without ignition source. If placed indoors, or in a closed gas cabinet, the consequences could have been serious.
3. The hydrogen discharged through a pressure relief device must be routed to a safe location, away from personnel, possibly at height (e.g. the roof of a building).
4. The logistic of the storage and supply system should facilitate quick and safe access to the shut-off valve of the bottle or cylinder.
5. Adequate ventilation is an important consideration in the layout of a compressed gas system.

This event touches upon a more general reflection on the need of maintenance and inspection of small pressure regulators. The manufacturer did not require a periodic maintenance program for the regulators of the type which failed. Nevertheless, they recommended the occasional replacement of the elastomer seal and o-ring on the outer part of the regulator body.
However, this recommendation cannot easily be translated in operative instructions, because ‘occasionally’ does not provide useful quantitative information on the lifetime of the seal. Moreover, regulators construction does not allow for easy dismantling and inspection actions, which bring with them the risk of dismantling/ re-assembling damages and errors. See a complementary discussion for a similar case of pressure regulator failure, HIAD_944, first reported by M. Henriksen et al, "Accidental hydrogen release in GC-laboratory; a case study", Int. Conference on Hydrogen safety, 2015 October 21-25, Yokohama, Japan.