Hydrogen explosion at an aerospace company

The event occurred in a support building at a test stand complex. The explosion destroyed two support buildings and severely damaged a large amount of equipment. No one was injured.

The high-pressure hydrogen gas system of the facility consisted of two storage tanks and connecting pipes. A location at one pipe located underground had been attacked by galvanic corrosion on the outer wall for several years, thinning the wall. However, the pipe was not routinely under pressure. Just before the explosion, an isolation valve failed open, permitting hydrogen under high pressure to enter the pipe and break the thinned wall. Hydrogen from the two 19.8 cubic meter (700 cubic foot) tanks was released in sand and gravel under the building, then entered the basement of the electrical control and instrumentation terminal building through ducts and throughputs. Eventually was transported through the air conditioning ducts from one support building into the next one.
Eventually, two large manual supply valves were closed, isolating the GH2 source about 30 minutes after the explosion. The facility's other supply systems and utilities had been severed or ruptured. Shrapnel and debris were ejected up to 540 feet away. Detectable levels of GH2 were recorded at several locations adjacent to the concrete pad for a five-day period following the event.

CAUSES
INITIATING CAUSE was the combination of two factors:
(1) a pipe thinned by corrosion and the failure of a high-pressure valve which accidentally released hydrogen into the damaged pipe. Galvanic corrosion had progressed at a pipe location where the prescribed protective coating was missing at the leak point. The original wall thickness of 15.2 mm (0.6 inch) had reduced to only 0.41 mm. The pipe was not in function and was not containing hydrogen.
(2) A pneumatically operated isolation gate valve, designed for 6000 psi service and located about 280 feet from the facility, failed in the open position. This was the results of ageing and degradation of the main seals. The failure caused hydrogen at a pressure higher than 4000 psi (275 bar) entering the pipe.

The IGNITION SOURCE was an electrical contact in a sump pump motor starter relay, which arced when the sump pump in the first building began an automatic start sequence. The electrical arc ignited the hydrogen in one building and this explosion apparently initiated an explosion in the other building.

ROOT CAUSES were (i) the wrong choice of the steel type for the piping, (ii) the lack of assessment of the extension of corrosion, (iii) lack of preventing safety measures, (iv) lack of inspection and maintenance procedures. In general, a lack of capacity by management to manage safety of the facility and managing changes.

isolation valve, storage tank, pipes

Pneumatic pressure had been removed earlier in the day and failure analysis indicated that the valve which failed had been damaged during recent field servicing. Leakage across the main seals of the valve over time, due to operations at another facility, permitted a pressure buildup within and downstream of the valve, sufficient to drive the valve to the open position.

DESCRIPTION OF THE COMPONENTS
The failed pipe was a carbon steel pipe with a 90 mm (3.5-inch) diameter and a 15.2 mm (0.6-inch) wall thickness. The pipe was coated with coal tar primer and coal tar enamel, wrapped with asbestos felt impregnated with coal tar, coated with a second coat of coal tar enamel, and wrapped in Kraft paper, a prescribed by a standard. The pipe was 8 ft 9 in below the concrete pad.
The failed valve was a hydrogen isolation gate valve, designed for 6000 psi service (415 bar). Pneumatic pressure had been removed earlier in the day and failure analysis indicated that the valve had been damaged during recent field servicing.
The two tanks had a volume of approximately 20 m3 each. By assuming a nominal pressure of 400 bar, a capacity of 500 kg of hydrogen each can be calculated. However, the hydrogen inventory when the release occurred is unknown.

No one was in the facility at the time of the explosion.
Damage significant, including destruction of two support buildings. Costs estimated approximately $5.9 million.

The event was witnessed about 30 seconds after the explosion by two employees 1000 feet away (30 m).
Authorities were notified and calls were placed to other personnel about 30 minutes after the explosion.
Teams of firefighters and emergency medical personnel were sent to the area to verify that no one was injured and to extinguish small residual fires.

[Lessons learnt simplified and elaborated from https://h2tools.org/sites/default/files/Hydrogen_Incident_Examples.pdf].

(1) Buildings near hydrogen systems should be designed to avoid the possibility of entrapment and accumulation of hydrogen (e.g., forced ventilation able to change volume of air according to the worse-case scenario, sloping roof, passive ventilation etc).
(2) It is better to avoid underground steel lines, especially if made of carbon steel. If unavoidable, transmission pipelines underground should be proof-tested and leak-checked on a periodic basis.
(3) Redundancy on isolation valves is required, including manually-operated ones. If remote-operated valves are required for operational purposes, they should be in series with and downstream of a manual isolation valve.
(4) The pressure between isolation valves and stand shut-off valves should be routinely and frequently monitored.
(5) All high-pressure gas lines scheduled to be inactive for periods greater than 6 months should be physically isolated by blind flanges from active systems.
(6) All reservoirs should be isolated at close of business each day, and before weekends and holidays.
(7) Corrosion protection systems for underground lines should be reviewed and tested to confirm the adequacy of the systems.
(8) Explosive gas detection meters should be included in the equipment carried by firefighters and emergency medical personnel.
(9) Fire alarm transmitters should be located at all hazardous locations.
(10) Emergency instructions should be permanently posted with names and telephone numbers of key individuals to be contacted.

According to (Cadwallader and Herring, 1999), "after that event, no mild steel was again used for high pressure hydrogen piping at that site".