LH2 release during transfer at an aerospace company
A liquid hydrogen leak occurred at the aerospace company (high Tier Seveso). The facility is specialised in production and handling of cryogenic propellants for engines of space vehicles. A 53 m³ liquid hydrogen road tanker had just been unloaded, its bottom valve closed, and the transferring hoses were being purged, when the portable gas detectors of the personnel on site triggered an alarm, indicating a hydrogen concentration in the air above 10% of the lower explosive limit (LEL). At the same time, operators observed a flow of liquid from a hose and white plumes of vapor emanating from the rear of the tank. The bottom valve of the tanker was not completely closed, resulting in a leak of liquid hydrogen upstream and downstream, particularly in the flexible hose, causing a flow along the surface of the hose due to the liquefaction of ambient air in contact with its uninsulated, cold surface. A second effect was a higher than usual release rate of hydrogen through the vent stack of the unloading area, leading to the detection of hydrogen by portable detectors. The detection was facilitated by the prevailing meteorological conditions, because the wind was blowing the hydrogen emissions from the stack towards the operators. The white plumes observed behind the tank were caused by the atmospheric conditions that day, which was cold and humid with a low dew point, favouring the vaporisation of liquid air and the condensation of air moisture. The emergency stop was activated, the facility shut down for safety and the area sprinkled with water. To bring the tank back to a safe state, the operator: (1) Opened the automatic valves of the tanker to direct the hydrogen to the main safety flare. This reduced the pressure in the road tanker and thus decreased the leak flow rate upstream / downstream through the incorrectly closed bottom valve. (2) Injected helium downstream of the bottom valve to dilute the hydrogen release and to warm the valve. This allowed the bottom valve to be properly closed and the facility to be returned to a safe configuration. Once the operations were completed, the emergency plan was deactivated approximately 7 hours later. Nobody was injured.

The INITIATING CAUSE was the only partial closure of a tank valve, which leaked liquid hydrogen once the hose was disconnected. The bottom valve of the tank had only partially closed. There valve had probably experienced an unexpected low temperature during the unloading process, leading to the formation of an ice plug of unknown composition, causing a mechanical blockage. This may have been exacerbated by a loss of vacuum in the double-walled enclosure of the valve or the piping to which it was attached. The ROOT CAUSE was probably a material failure, but a mistake in the execution of the transfer procedure cannot be excluded either.