Explosion at a filling hose of a hydrogen dispenser

The incident occurred at a dispenser of a hydrogen refuelling station (HRS).
The HRS employee had just terminated the filling of a fuel cell vehicle (FHV) and was depressurising the filling hose, when a gas detector near the filling nozzle detected a leak and issued a visual alarm.
However, the employee was in the blind spot of the alarm light and did not notice it.
After depressurising the filling hose, the employee removed the filling nozzle from the FCV and stored it in the dispenser. The filling hose exploded with an explosive sound, breaking open the thermoplastic resin outer layer of the filling hose.
No personal injury occurred.

The following is a chronological summary of the accident:
• 3:45 PM - The FCV arrived at the HRS.
• 3:48 PM - The FCV began filling with hydrogen.
• 15:52, the hydrogen filling into the FCV was terminated. While the filling hose was being depressurized, the suction-type gas detector near the filling nozzle triggered an alarm at 500 ppm of hydrogen. One second later, a high alarm at 1,000 ppm was triggered and the hydrogen compressor automatically shut down.
• The employee was in the blind spot of the rotating beacon (patrol light) installed near the dispenser, so they did not notice the alarm.
• After depressurisation of the filling hose, when the employee removed the filling nozzle from the FCV receptacle and placed it back into the dispenser, a loud pop was heard and the filling hose ruptured, scattering the thermoplastic resin outer layer of the filling hose.

INITIATING CAUSE
The inner layer of the filling hose failed due to material fatigue under harsh operative conditions: ultra-low temperatures of approximately -40°C) and repeated pressurisation and decompression cycles from 0 MPa to 82 MPa.
The investigation revealed that a small hole developed in the innermost layer (resin, approximately 1 mm thick) near the nozzle end of the filling hose.
The hydrogen leaked then through the hole into the reinforcing layer, forming a gas reservoir between the innermost and outer layers along the entire length of the hose (3,500 mm).
When the filling hose was removed from the vehicle's receptacle and returned to the dispenser, the outer layer ruptured, causing the hydrogen to scatter.

ROOT CAUSE(s)
This filling hose was an older model, a prototype for demonstration testing, and was not as durable as the new model with an improved innermost layer of material and a longer lifetime.
The manufacturer's warranty period for this filling hose was one year after the start of the operation or 500 pressurisations, whatever came first. However, because it had only been pressurised 330 times, it was used for a period exceeding the manufacturer's warranty (one year and four months).
This use a component beyond its guaranteed lifetime was a management decision. An additional element of the root cause was also the loose relationship between the simplified tests executed to assess/verify hose performance and the real (harsher) operative conditions.
Finally, the failing to realise that an alarm had been triggered and that the dispenser had been shutdown is an indication that the design of an emergency situation needed improvement.

hose

DESCRIPTION OF THE FACILITY
The hydrogen refuelling station had a production capacity: 30,606 m³/day of hydrogen delivered at a nominal pressure 82MPa and a nominal temperature -40°C
The material of the innermost layer was thermoplastic resin with an approximate outer diameter of 14.8mm and a length of 3.5 m.

No personal injury occurred. or property damage, beyond the need to replace the filing hose

[Fundamental Lessons Learned]
This incident, as well as similar ones which occurred around the same period (see HIAD no. 1196) highlighted the lack of correlation between the design verification test and the real operative conditions. The verification by means of the "hydrogen impulse” test " used for the development and production of the filling hoses appeared not able to catch the real degradation occurring during the pressurisation - depressurisation cycles at 82 MPa and -40 C at the HRS. The consequence was that cracks developed in the innermost layer of the filling hose, leading to leaks, well before the number of fills recommended by the filling hose manufacturer was reached. It was recommended that hose manufacturers would keep investigating the various degradation mechanisms, to be able to reflect them in a more realistic performance assessment test.
[Lessons Learned in Development]
A better quality (control) of the polymer material used in the innermost layer of the filling hose could help in improving uniformity and in preventing crack initiation.
[Lessons Learned for Hydrogen Station Operators]
Filling hoses must be used within the manufacturer's warranty, and must be replaced promptly if the period of use or number of pressurisations reaches the warranty limit.

• The filling hose manufacturer replaced the hose with an improved model featuring an improved innermost layer, designed to withstand more repeated pressurisation and decompression in an ultra-low temperature, ultra-high pressure compressed gas environment.
• In cooperation with the filling hose manufacturer, the basis for the manufacturer's warranty period and number of uses for the filling hose were reviewed and re-verified, to ensure that the harsh operating conditions such as the 82 MPa filling and continuous filling were reflected in the design performance tests.
• It was decided to execute regular endoscopic inspections of filling hoses.
• Additional rotating beacons (patrol lights) were installed, to ensure that employees realise the triggering of an alarm, even while they are busy in filling an FCV.