The successful outcome of the liquid-metal leak test was a key event in the MEGAPIE (MEGAwatt PIlot Target Experiment) project, a multinational endeavor aimed at developing a reliable neutron spallation source operating with dense liquid metal. Indeed, the leak test validated the containment design, which was a regulatory requirement for demonstrating that a liquid-metal source could be operated safely. Furthermore, unique temperature and stress measurements were recorded that agreed well with test predictions published ahead of the test. This paper outlines the approach taken for predicting the consequences of a liquid-metal leak, with particular emphasis on a simplified one-phase calculation method that may be useful in the future for predicting the impact of accidental liquid-metal leaks at modest expense in terms of CPU time.

Most of the assumptions underpinning the original analytical predictions necessarily erred on the conservative side. Therefore, the boundary conditions applied to the original analysis, such as the exit flow rate of the liquid-metal jet, are critically reviewed in this paper to improve on the existing agreement between the predictions and the experimental data.