Incident Synopsis
While attempting to replace a rupture disk in a liquid H2 vessel, H2 gas was released and ignited. In fighting the fire, liquid N2 was sprayed onto a second liquid H2 vessel located nearby. This resulted in cracking of the outer mild steel vacuum jacket. The loss of the vacuum caused a rapid increase in pressure and rupture of the burst disk of the second vessel. H2 boiled off and was burned in the fire.
Cause
The rupture disk was being replaced with a load of liquid H2 in the vessel and no separating inerting gas. The H2-air mixture was probably ignited by static discharges. Rupture of the second vessel burst disk was caused by the low-temperature exposure of the mild steel vacuum jacket.
The use of inerting gas or other means of separation should be employed when conducting mechanical work where hydrogen gas could be present. More importantly, per CGA S1.3, the vessel should be equipped with a dual relief system that can isolate one side from the other and allow a rupture disc to be changed without exposing the operator to hydrogen.
In the second incident, the cracking of the outer mild steel vacuum jacket was more than likely related to the coefficient of thermal expansion of steel, which defines how much the material will contract when its temperature is decreased. The temperature of cryogenic liquid nitrogen is at -195.8 °C (-320.44 °F), and the linear coefficient of thermal expansion of 1020 steel at room temperature is 12 x10-6 1/0 °C. Thus, the significant contraction in the steel due to the instantaneous temperature reduction created localized stresses, which cracked under the vacuum pressure of the system. Some other method of controlling the fire should have been employed. In addition, the metal would have been made much more brittle due to the low temperature.
All relevant personnel should receive at least basic training on the proper selection of fire extinguishing techniques for the given scenarios they are likely to encounter.
Liquid nitrogen should not be used to put out a hydrogen fire. It is very difficult to put out a gaseous hydrogen fire, plus had the liquid nitrogen not cracked the nearby vessel's shell, it certainly could have cracked the original vessel. It could also have plugged up the stack by freezing at liquid hydrogen temperatures.