A Hydrogen-air Explosion in a Process Plant: A Case History

In the summer of 1985 a severe hydrogen-air explosion occurred in an ammonia plant in Norway. The accident resulted in two fatalities and the destruction of the building where the explosion took place. This paper presents the main findings from an investigation in 1985 and 1986 of the gas explosion and its consequences. The event started when a gasket in a water pump was blown out. The water pump was situated inside a 100 m long, 10 m wide, and 7 m high building. The pump was feeding water to a vessel containing hydrogen gas at pressure of 30 bars. This pressure caused a back flow of water flow through the pump and out through the failed gasket. The hydrogen reached the leakage point after about 3 minutes. The discharge of gas lasted some 20 to 30 seconds before the explosion occurred. The total mass of the hydrogen discharge was estimated at 10 to 20 kg hydrogen. The main explosion was very violent and it is likely that the gas cloud detonated. The ignition source was almost certainly a hot bearing. Several damage indicators were used to estimate the amount of hydrogen that exploded. The indicators include deflection of pipes and panels, distances traveled by fragments, and the distribution of glass breakage. We found that 3.5 to 7 kg of hydrogen must have been burning violently in the explosion. Window glass was broken up to 700 m from the centre of the explosion. Concrete blocks, originally part of the north wall of the building and weighing 1.2 metric tons were thrown up to 16 meters. The roof of the building was lifted by an estimated 1.5 meters before resettling. The displacement of the roof caused a guillotine break of a 350 mm diameter pipe connected to the vessel that was the source of the original gas discharge. The gas composition in the vessel was 65 - 95 %2hydrogen. This resulted in a large horizontal jet fire lasting about 30 seconds. Minor explosions occurred in the plant culvert system.To our knowledge this gas explosion is one of the largest industrial hydrogen explosions reported. We believe this case history is a valuable reference for those who are investigating the nature of accidental hydrogen explosions.