HSE investigations into the brittle fracture of the outer jacket of a 60 000 litre liquid nitrogen tank have found that piping hidden inside the vacuum interspace between the inner vessel and outer jacket developed a fatigue crack after approximately nine years operation. Loss of containment from the economiser pipework resulted in impingement of cryogenic liquid  / gaseous nitrogen on to the carbon steel outer jacket which subsequently suffered a brittle failure at a low pressure ejecting steel fragments over the surrounding area. The outer jacket material is carbon steel and inner vessel and piping austenitic stainless steel. Another HSE investigation of the same economiser piping assembly in another tank of the same age and operating conditions revealed the start of a similar fatigue crack. One Cryogenic tank manufacturer reported eight cases of cracks on an earlier economiser piping design within the vacuum interspace. At the time these were attributed to pipe/weld failure.
On such tanks, thermal expansion and contraction of pipework within a vacuum interspace is accommodated by adding a horizontal (expansion) leg to any vertical pipe. There is some evidence that on the tank that suffered the failure, the movement of one of the horizontal legs may have become constrained. Subsequent pipe dimensional changes, generated by temperature changes over the significant length of pipe, resulted in unacceptable stress levels at the welded joint between two pipes at the top of the tank. Continued high cyclic loading resulted in the generation of a fatigue crack at the toe of this piping fillet weld.
In the first failure investigated, an initial small cryogenic leakage through the fatigue crack agitated the surrounding perlite, resulting in erosion through a section of a neighbouring pipe and a much larger release of cryogenic fluid.
HSE understands that constraint of the lower horizontal leg may have been due to compaction of the interspace perlite powder insulation as the horizontal pipe section moved up and down during repeated pipework thermal expansion and contraction cycles. Further investigations into this mechanism are ongoing and the results may have implications for other forms of insulation or for the more general case of pipe movements in fluids denser than air.
The number of cycles to failure is difficult to predict and will vary with the degree of constraint on the pipework, the particular pipework geometry, the temperature differentials experienced, customer demand patterns and the quality of the stressed weld.
It is not known how long the period can be between a through wall crack appearing at the stressed weld and sufficient cryogenic fluid entering the vacuum space to cool the outer jacket to a temperature that would allow brittle failure. Failures are expected to follow a defined pattern of; loss of vacuum, operation of vacuum relief device(s), localised icing of the jacket near to the leak and then brittle failure.
Recognition of a leak within the interspace at a very early stage is expected to give sufficient time for the tank to be made safe.
Cryogenic liquid storage tanks have been run safely in the UK for many years with no previous significant incidents reported.
Cryogenic liquid storage tanks with these economiser piping assemblies can, under some operating configurations, be subject to repeated and significant differential thermal expansions within piping systems that increase the risk of a hidden failure.
Impingement of cryogenic fluids onto the carbon steel outer jacket of a cryogenic storage tank could lead to brittle fracture and the ejection of fragmented steel. Such ejected fragments could cause serious injury (including death) to individuals in the vicinity of the tank, and they could also compromise nearby safety critical plant or containment systems such as pipes, tanks or other vessels.