Piping failure at a nuclear power plant due to hydrogen explosion

A hydrogen explosion ruptured and fragmented a 6-inch steam condensate pipe elbow at of a nuclear plant. The component affected was a steam piping of a heat exchanger used for removing residual heat from the reactor. The event occurred while the plant was in operation, during a test of a high-pressure core injection system for cooling water used for emergency cooling.

The piping was turned up from inside and was completely broken. The cause of the explosion was believed to be the accumulation and detonation of hydrogen gases created by radiolysis of water or steam. The ignition source was not conclusively identified, although some sources advanced the hypothesis that a ignited spontaneously assisted by the presence of a platinum particles in the system, which acted as catalyst.

The incident resulted in a leakage of radioactive gases. The part of the piping broken was located on the 2nd floor of the building. No worker was there, but there were 36 workers in the radioactivity control area inside the building.

The INITIATING CAUSE was the formation of an explosive hydrogen-oxygen mixture in a section of a water vapour pipe.

Although originally some sources advanced the hypothesis that the IGNITION SOURCE could have been the presence of a platinum particles in the system, later studies demonstrated that the explosive mixture can auto ignite under sever fluid pressure transients.
Moreover, the pipeline affected by the explosion had been modified 7 years before the incident: the new line was longer and characterised by a higher number of bends and curves, which can act as accelerators of pressure waves. This could explain why this event never occurred I the long operative history of the plant.

The ROOT CAUSE was the lack of knowledge of the phenomenon and he obvious related hazards. When designing and operating a system containing water steam, it was assumed that no non-condensable gases (hydrogen, oxygen, nitrogen, etc.) could exist in its pipes.
Moreover, with a pressure of 7 MPa and a temperature of 270 C, it was assumed that even in the case of formation of explosive mixture, the conditions were too low for its autoignition. The possibility of a situation like that created by a ‘water hammer’ effect had not been considered.

heat exchanger, pipe

The leak occurred during a test aiming at assessing the performance of a high pressure core injection system

The BWR plant had a rated power of 540 Mwe, it started the operation in 1976.
The ruptured pipe had a inside diameter of 15 cm and a wall thickness of 1.1 cm.

Property losses are unknown but were probably limited to the replacement of cooling system affected by the explosion, and the costs of decontaminating the affected area.

Despite the local emission of radioactivity, this incident had no impact on external areas of the station, as safety systems functioned adequately.

However, it was an event not recorded before in Japan. Moreover, it was soon followed by a similar event at a German BWR plant (14 December 2001, see HIAD_522). The two events highlighted the urgency to assess the risks associated with the accumulation and explosion of hydrogen by radiolysis in BWR.
Since many studies were performed to identify the conditions under which oxygen and hydrogen generated by electrolysis come to an explosive event inside pipes.
The analysis of the event performed by Naitoh et al. (2012 see References) concluded that in the section of the pipe which exploded in this event, a non-condensed gas atmosphere had formed consisting mainly of H2 and O2 because of factors such as the pipe geometry (a series of vertical and horizontal sections) and the temperature and heat experienced by the water vapour.
According to A. Leishear (2020, https://doi.org/10.1115/1.4044807 , see References), fluid transients are then responsible to heat the explosive gas mixtures to autoignition. The same mechanism was responsible for pipe and pump damages to U.S. reactor systems since the 1950s, while at that time water hammer alone had been assumed as cause.

Several studies based on computational fluid dynamics studies coupled to stress analyses have since informed designers and operators to avoid the occurrence of the conditions bringing to such events.
Already in 2006, Antaki et al (see References) prosed the following approach to assess the risks associated with the accumulation and explosion of hydrogen generated by radiolysis:
(1) To identify the systems susceptible to hydrogen accumulation, and their initial conditions (hydrogen-oxygen mixtures, their location and their initial pressures and temperatures).
(2) To determine whether the explosion is a deflagration, a detonation or a deflagration transitioning into a detonation
(3) Based on the established conditions and the explosion regime, to develop the explosion pressure timelines and peak temperatures. This is the fluid dynamics aspect of the assessment.
(4) To calculate the structural capacity of systems, equipment and components subject to the calculated explosion loads.
(5) To perform the stress-strain analysis and fracture analysis.
(6) If the results of (5) shows a situation exceeding the structural integrity capacity of the structure, to investigate the failure mode to understand the consequences of such an explosion.