rupture LH2 tank

The incident occurred on a large LH2 stationary storage tank at an aerospace testing facility. The cryogeninc tank experienced a significant vacuum leak in the insulation annulus, which eventually led to major cracks in the bottom of the outer shell.

Two rupture disc assemblies at the top of the tank developed leaks and a pressure rise was noted in the annular space. However, due to the need to perform incumbent rocket engine test, it was decided to continue the operation of the tank instead of removing it from service. A normal liquid top-off induced a cryo-pumping resulting in the drop of the annular space pressure from approximately 9332.6 Pa to 3.1 Pa. During test runs, operational constraints were implemented to prevent shifting and damage of the inner vessel. After completion of the test, the tank was drained to facilitate repair of the leaking rupture disk assemblies. However, two days after the tank had been completely emptied, heavy frost developed on the bottom of the outer vessel. Temperature sensors indicated the outer tank wall was less than 90 K at the bottom, which is well below the ductility range for its carbon steel outer wall. Additional vacuum pumping was attempted but proved difficult due to small port sizes and/or perlite intrusion into the pumping system. Eighteen days after tank drain, the outer vessel cracked.

The INITIATING CAUSE was the development of cracks at joints in the outer shell, which caused air entrance into the insulating vacuum shell.
The ROOT CAUSE was manufacturing error and the choice of low strength steel for the material of the outer shell.

DESCRIPTION OF THE SYSTEM
The tank vertical cylindrical structure with a volume of 340.7 m3 (a capacity of 25 metric tons of hydrogen). It was built in 1962 and it outer was made of carbon still.

This event was a near miss. The operation of the tank was stopped short before the failure cracking of the external shell occurred. Despite the presence of safety measures such as pressure relief devices, the sudden loss of thermal insulation would have certainly caused a massive boil off rate.

A general lesson learnt from this event and similar cases was drawn by Krenn in his scientific article on “Annular air leaks in a liquid hydrogen storage tank” (see reference), where also diagnostics and mitigating solutions are proposed:

Entrance of air in the annulus can be very difficult to identify in a timely manner, especially if the air flow is small. Contrary to room temperature vacuum leak, easy to detect due to increase in pressure, the increase in pressure in the annulus due air inflow, the extremely low temperatures cause most of the air constituents to freeze near the inner tank wall. Only residual helium and neon in the air contribute to a pressure rise, which may be so slight that it goes unnoticed for months, or even years. Often the problem is first identified due to increase in boil-off. As air freezes inside the perlite, the thermal conductivity of the system increases due to thermal shorts and residual gases. The heat leak can be significant enough to cause a noticeable surge in the boil off rate.

To minimize fabrication costs, LH2 tanks had often an outer wall made of carbon steel, which is highly susceptible to corrosion and minor defects may grow into leaks. Soft seals, at tank joints, may also develop leaks over time, as it happened in this event.

Long term monitoring of the annular pressure can identify a leak and may even provide an estimate of the leak rate. This is possible because the fractional content of helium and neon in air is well known. Therefore, once the annular pressure rise is noticeable, a gas analyser could measure the gas composition of the annulus, and from the concentration of helium and neon and the vacuum value deduce the amount of air entered/entering the annulus.