Hydrogen explosion at renewable hydrogen production facility

The hydrogen buffer tanks that exploded were part of an experimental facility experimenting generation of renewable hydrogen from a water electrolyser coupled to solar panels.
The three hydrogen tanks (40 m3 capacity each at pressures of 1,2 MPa one of them and 0,7 MPa the two others) were receiving the hydrogen produced by the electrolyser.
Firefighters said all three tanks were destroyed in the explosion; there was no fire, but the explosion was strong enough to send debris being scattered in an area well over 3,000 square meters. "It's not known whether three tanks exploded at the same time or whether it was a series of explosions".

The equivalent TNT is estimated to be of about 50 Kg based on the damage near by. There was no secondary fire and the damage is merely by the detonation pressure. The explosion killed two people, injured six others, and caused damage to a nearby ceramic components manufacturing plant. The buildings within 100 m from the tanks were heavy damaged. People in the city at 6-7 km from the site heard the explosion.

According to the preliminary results of the legal investigation (still ongoing, October 2020) and no final report about the causes is yet available (in Korean, translated into English by a non-official translator), the hydrogen and buffer tanks exploded due to a static spark in the hydrogen buffer tank, with oxygen concentration exceeding 6%, which is explosion threshold.
The presence of oxygen in the hydrogen stream was due to a wrong operation of the water electrolyser, which had been operated at a power lower than the minimal power level. Due to the asbestos type of separation membrane adopted in the electrolyser, the system has to be operated above the 98 kWh threshold, which corresponded to 50% of the nominal power. On the contrary, due to characteristics of the solar panels providing the power, the system often had often operated at a power lower than this threshold.
The investigation has also identified a series of contributing causes:
(1) The option to use an oxygen removing component was not considered in the final design, as well las the static spark remover in the hydrogen buffer tank. These decisions were probably related to the need to reduce costs.
(2) Oxygen concentrations were detected higher than 3%, but it was decided to continue operation because to achieve the 1000 hours of operation necessary to validate the tests.
(3) The safety management team did not followed the safety requirement to daily test the hydrogen quality.

water electrolyser, hydrogen storage tank

The event occurred during a validation test of the whole facility, aiming at demonstrating its continuous operation for 1000 hours.

DESCRIPTION OF THE FACILITY
The plant was a demonstration site for the generation and utilisation of renewable hydrogen, consisting of photovoltaic panel, an electrolyser, storage tank and a fuel cells system.
The 200 kW alkaline electrolyser had a capacity of 40 Nm3/hr and delivered hydrogen at 1.2 MPa, which was stored at the same pressure in a 'buffer tank' of 40 m3 storage capacity (ID: 2,450 mm, H: 7,668 mm). After the buffer tank the gas was delivered to two additional tanks, probably of the same type as the buffer one, but kept at 0.7 MPa. These two were delivering the hydrogen to a fuel cell system for power generation.

The explosion killed two people, injured six others, and caused damage to a nearby ceramic components manufacturing plant. The building within 100 m from the tanks were heavy damaged. People in the city at 6-7 km from the site heard the explosion.

The Korean Supreme Court ordered over 10 billion won [$6.8 million USD] in compensation from five organizations for the explosion. The court judged that the organisations had competences and capacities to foresee the accidetnal situation, but failed to install devices such as rectifiers, oxygen monitors, and oxygen removers which would have been able to prevent prevent it. Amon gother findings, the court identified clear negligence in improperly evaluating and managing project contractors.

The accident is still being investigated and no final report about the causes is yet available. A preliminary technical analysis has been made publicly available (see references).
According to this preliminary analysis, the plant showed design deficiency (lack of preventing and mitigating measures) and was operated at out-of-specification conditions.
These shortcomings need to be corrected, for example by providing in-situ diagnostic system able to trigger emergency stops of the hydrogen production system and an automatic isolation of the storage.

The most plausible initiating cause was the defective functioning of the electrolyser’s membrane, at lower power, which caused oxygen diffusing into the hydrogen stream. If this is confirmed, the following corrective actions are required to fill a global knowledge gap:
(1) better understand the relationship between the gas permeability of the electrolyser membrane and dynamic operation range caused by the variability of renewable power sources (solar);
(2) Improvement of standardised performance and safety tests, aiming at defining a more realistic testing requirements and conditions at partial/low load cycles. This may bring to a new version of the ISO 22734:2019 "Hydrogen generators using water electrolysis".