Failure of a pipe a ammonia plant due to hydrogen embrittlement

The incident happened in an ammonia plant, on the pipe leading hot synthesis gas (syngas) from the synthesis convertor to the first heat exchanger. The pipe cracked, releasing the syngas which produced a small fire.

This pipe had already experienced six previous failures and consequent leakages. Following each cracking, repair/re-welding was performed. Before the last cracking, part of the pipe had been replaced with a new piece. All cracks occurred at the same location, an approximately 90 degrees bend of the pipe.

The scientific article used as source (see references) reports that the design and the material adopted for the pipe was correct, and the pipe was operated within the recommended temperature and hydrogen partial pressure limits. Most of the cracks were due to hydrogen embrittlement and/or insufficient heat treatment in the welds and their heat affected zones. Additional causes were additional local microstructural aspects contributing to degrading the mechanical performance of the bend in respect to the design values assumed for the base material.

The INITIATING CAUSE was the cracking of a pipe and the consequent release of syngas.

The referred scientific article reported the following findings:
1. All six failures could be attributed to incorrect heat treatment.
2. Their modified microstructure failed due to hydrogen embrittlement.
3. The geometric deformation in the bending caused a doubling of the stresses of the straight (and perfectly round) pipe
The ROOT CAUSE could be then identified in inadequate design and component manufacturing.

syngas pipe

This pipe had been affected by six cracks and consequent leakages, along a time span from 1978 to 1982. Following each cracking, repair/re-welding had been performed. Before the last cracking, part of the pipe was replaced with a new piece.

According to the referred source, the pipe was constructed from a hydrogen embrittlement resistant steel. Operating conditions were 380°C at 320 bars.

The plant had started operation in 1970

This case highlights the difficulty to guarantee the correct material specifications (composition, micro-structure, heat treatments, in-situ modification) required to avoid hydrogen embrittlement in large-dimension component.

The plant operator adopted a very careful procedure for the repair of cracks (see for details the scientific article used as source):
(1) The bends were jacketed and provided with nitrogen protection. The nitrogen injected at one
end of the jacket was analysed regularly at the exit end for the presence of synthesis gas.
(2) Ultrasonic testing was carried out on the welds and bends on the rare occasions that this was possible.
(3) No rapid changes of pressure or temperature during running.
(4) The pipe was pressure tested with hot water at design pressure after each weld repair.