PRD failing to open

A cryogenic hydrogen laboratory had a power outage. The cryogenic hydrogen storage was no longer being cooled, therefore temperatures and pressures within the storage system started rising. There were several relief devices on the system, one with a set point of 150 psi (approximately 1 MPa) and a second at a set point of 165 psi (1.1 MPa). The system was monitored by site personnel so that when the hydrogen pressure increased to 120 psi (approximately 0.5 MPa), a manual vent was opened. However, the manual vent rate was not able to control the rising pressure. Pressure continued to increase until a 165 psi relief valve opened to control the pressure. The 150 psi relief valve failed to open. This incident was classified as a near miss.
[Source: Burgess et al., Technical Report NREL/TP-5400-67381]

The INITIATING CAUSE was a power outage, which stopped the cooling of the cryogenic storage.

Inspection of the failed relief valve with a 150 psi set point showed an opening pressure of 230 psi (approximately 1.6 MPa, 50% above set point). It is suspected that vacuum grease solidified (froze) in the valve seat. In this case the combined manual venting and secondary 165 psi relief venting were sufficient to control the overpressure condition. The 150 psi set point valve was in a section of the system that was not expected to see a cryogenic temperature so was not rated for cryogenic service.

The ROOT CAUSE was the installation of a relief valve not suitable for cryogenic temperature, because its location was not expected to be affected by these temperatures.

cryogening tests, PRD, vent stck

A power outage was at the start of this incident .

Despite the malfunctioning of one of the two pressure relief valves, the combined manual venting and the secondary 165 psi relief venting were sufficient to control the overpressure condition. Therefore the event is labelled as near-miss

Despite the emergency caused by the loss of power and coolant, the system affected by the liquid spill was equipped with redundant safety system, so that consequence of the loss of containment were limited and no escalation occurred. Indeed, the investigation concluded that the combined manual venting and secondary 165 psi relief venting were sufficient to control the overpressure condition.
The NREL report concluded that pressure relief devices should be put on a predictive maintenance schedule that depends on the severity of service. Further information on predictive maintenance and failure mode determination can be found in the American Petroleum Institute’s Recommended Practice RP-576 (American Petroleum Institute 2009). Moreover, relief valves for liquid hydrogen system must be valve rated for hydrogen at the related temperatures range. The design improvement allows for migration of cryogenic boil off to the relief system.
Additional return of experience can be found in the scientific paper of a. Jimenz and K. Groth:
(1) Not correctly understanding the risks and hazards led to installing the cryocooler and fume hood without backup power.
(2) Asset integrity, quality control (e.g., design, installation, procurement), and operational readiness should have identified that the valve had the drainage weep port and removed it for use. These programs should verify that the valve is appropriately set and specified before use.
(3) The asset integrity program should have tracked regular tests and calibration to ensure operations.

The 150 psig relief valve, which failed, was replaced with a valve rated for cryogenic hydrogen. The design improvement allowed the correct flow of boil-off hysdrogen to the relief system.