Release of hydrogen from a LH2 storage tank

A hydrogen release occurred at hydrogen storage unit of a manufacturing plant in an urban area.
The storage unit was a 9,000-gallon (34,069-liter) liquid hydrogen storage vessel.
Its internal pressure increasing beyond the maximal allowed value and a safety relief device (a burst disk) activated releasing hydrogen into the atmosphere through the vent stack. The hydrogen jet rose approximately 15-20 feet (4.6-6.1 meters) in air, but it did not ignite.

The industrial gas company responsible for the cryogenic installation was notified. They sent service technician who found the tank pressure at zero and the burst disk blown, switched the three-way diverter valve to the other safety relief device and replaced the burst disk which had activated.

There were no injuries or damage from this incident, the hydrogen safety venting equipment functioned properly. The hydrogen was released without ignition, and since storage was in open space, no cumulation of gas in closed space occurred/ Normal operations could resume after it was determined that there were no unsafe conditions.

However, approximately seven months later, the burst disk that had been replaced in the first incident ruptured for the same reasons. No outside emergency response was involved with this second hydrogen venting incident, as once again the pressure relief system worked as designed.

The INITIATING CAUSE was the correct activation of a burst disk due to overpressure.

The cryogenic storage vessel had not been used for a (unspecified) long period, so that temperature gradient from ambient air temperature (60°F/16°C) through the vessel vacuum-jacketed walls were able to rise the internal temperature. The normal storage pressure for this vessel was 150 psi (10.3 bar). A factor contributing to a premature activation was the existence of a backpressure against the disk.

The ROOT CAUSES could be a combination of a lack of regular and effective inspection and of a inadequate venting design related to the existence of a backpressure of the burst disk. However, regarding this latter point, see diuscssion udner Lesson Learnt).

bursk disc, piping

The cryogenic storage vessel had not been used for a (unspecified) long period

The only property loss was due to the loss of hydrogen.

The hydrogen venting functioned properly. The hydrogen was released without ignition, and since storage was in open space, no cumulation of gas in closed space occurred.
Nevertheless, the following recommendation should be issued:
1.The emergency revealed that local emergency responders were not well informed on hydrogen behaviour and aspects of hydrogen safety required to assess the situation and to intervene safely and effectively. Specific training should be provided.
2. Liquid hydrogen installations even if they are not used, especially if they contain a non-negligible quantity of hydrogen, should be inspected by facility personnel on a regular basis, to verify proper operation and identify physical damage or leaks. In the US, the inspection should be consistent with national Fire Protection Association Code NFPA 55. In case of deviations, the site operator should contact the servicing company.
3. In the US, the Compressed Gas Association (CGA) publications has published a series of guidelines and standards for the design, installation and maintenance of liquid hydrogen installations: G-5.4 - Standard for Hydrogen Piping Systems at Consumer Locations, G-5.5 - Hydrogen Vent Systems, H-3 - Cryogenic Hydrogen Storage, and H-5 - Installation Standards for Bulk Hydrogen Supply Systems.

[Note of HIAD Validator: the source (H2TOOLS) identifies as technical lesson that burst disks are highly sensitive to any form of back pressure. Although true in general, it is an unclear point for this specific case, i which the pressure was raising internally due to increased heat. The report does not provide enough details, such as the values of the design and real pressure downstream of the disk and the reason for a change of this pressure. On one side, it is said that the activation of the burst disc was pre-mature, i.e. if failed before its nominal activation pressure value. On the other side, the report attributes to a backpressure the reason for the pre-mature activation. A burst disc works on a differential pressure principle, any increase in backpressure will increase the activation pressure, causing the disk to activate , later, not earlier than designed. The only way in which a disc can fail prematurely due to a change in backpressure is a case in which that pressure decreases in respect to the design value. How this could happen is not explained. ]

Following this incident, some modifications were made to the hydrogen piping to eliminate all back pressure on the burst disk.