Several leaks from the hoses of a HRS

This event summarises 3 very similar cases of leaks at dispenser hoses during refuelling, occurred at the same hydrogen refuelling station in the period January – April 2017.
The leaks triggered an automatic or manual shutdown and were also noticed by their hissing sound. In all three cases, the leaks were caused by the degradation occurring at a number of fillings much lower than the one indicated by the hoses manufacturer as a typical lifetime. Despite difference in the detailed mechanism of the degradation, all three cases were characterised by the formation of a crack at the internal surface of the most internal layer of the hoses, followed by crack growth along the pressurisation-depressurisation cycles. In two cases, a stating damage was caused by small spurious particle present in the hydrogen flow.
The quantity of hydrogen released were not reported, but were classified as “other than trace amount”.

The INITATING CAUSE was the failure of the internal layer of the hoses
The ROOT CAUSE was a shortcoming in the performance of the hoses, and in particularly of the resistance to hydrogen and mechanical fatigues experienced by the most internal layer of the hoses. The damage induced by small metallic particles indicate also shortcoming in the installation phase of the hoses.

hose

DESCRIPTION OF THE FACILITY
The HRS was a 70MPa refuelling station with a capacity of 34005.3 m3/day at 82 MPa, with pre--cooled hydrogen delivery.

in all cases the finacial damage consisted costs of new hoses and the failed delivery of hydrogen to customers.

(this is a strongly reduced version of the original KHK report). These hose failures highlighted the need for better materials, better designs and better performance tests.

[Fundamental lessons]
The hoses failed by internal cracks well before the indicated lifetime by the manufacturers. It appears that the test used during hose development to verify the lifetime of hoses based on the number of "hydrogen impulse testing" was not able to represent real operative conditions at the dispenser. There was the need for hose manufacturers to perform a reality check at the at hydrogen refuelling stations, to identify and understand the degradation mechanisms leading to cracks, and to use the findings to improve the testing methods.

[Lessons learned in development]
It was also clear that the polymer material used in the innermost layer of the filling hose needed improvement to prevent occurrence of blisters, which could represent the starting point for cracks.
Moreover, the overall structure of the reinforcing layer required improvement to better prevent and reduce distortion of the filling hose. Since circumferential strain affects fatigue failure of the innermost layer, it emerged the need to reduce this strain by increasing the braid angle of the special high-strength fibre in the reinforcing layer. However, care must be taken to avoid reducing pressure resistance and durability of the filling hose, so that an additional steel wire reinforcing layer could be required.

[Lessons learned when installing and replacing hoses]
Filling hoses through which low-temperature, high-pressure hydrogen passes can develop cracks due to the impact of small amounts of metal powder remaining in the flow path, and the subsequent cycle of pressurization and depressurisation can cause the cracks to grow and penetrate the hose. When installing or replacing a filling hose, it is important to take care to prevent metal powder from remaining in the flow path, such as by purging the flow path with nitrogen.

The first measure was obviously the replacement of the filling hoses new ones.
The operator of the HRS affected changed the filling hose manufacturer. In addition, an information campaign started to make aware other hydrogen stations using filling hoses of the same manufacturer on how to manage the hoses.

At the manufacturing plant, foreign matter has to be removed by air blowing (flushing) at after each production process and before shipping, all filling hoses must be inspected endoscopically to confirm absence of foreign objects. Rubber caps must protect both ends and covered with vinyl.

When assembling on-site, flushing with nitrogen gas at 1.5 MPa to eliminate particles (including the decompression line) and their absence must be checked by using a white cloth.
The frequency of replacing the filling hose was increased.
The filter mesh size was changed from 10µm to 2µm.