The spray safety system is one of the most effective solutions to limit the pressure and temperature rise in the containment of a nuclear power plant during a design basis accident. However, its actuation under severe accident conditions may have either a positive effect on the hydrogen risk by mixing the containment atmosphere and avoiding high H2 local concentrations or a negative impact by condensing the steam and de-inerting the atmosphere. The experiments performed during the OECD/NEA HYMERES-2 Project aimed to extend the experimental database dedicated to assessing the capabilities of the codes used to optimise the operation of the safety systems. Specifically, the H2P5 series consisted of two experiments that compared the activation of a central full cone spraying nozzle (H2P5_2) with nine smaller spraying nozzles, eight equally distributed at 0.5 m from the metallic walls of PANDA (H2P5_1). The ninth nozzle was positioned in the vessel vertical axis for the latter configuration. H2P5_1 was the first experiment ever using a multiple nozzle configuration in PANDA, which was aimed to address the conditions expected in a plant containment with droplets interacting with the vessel walls as well as with each other.
The measurements indicated a considerable difference between the two configurations regarding the pressure and temperature decay, which signified the presence of non-adiabatic effects. The stronger cooling of the walls measured during H2P5_1 resulted in a slower depressurisation for most of the test duration. This can be explained by an intensified interaction of the injected droplets (by direct droplet-wall interactions or by potential formations of liquid films) and the vessel walls, which was coherent with the use of the spray ring in H2P5_1. It was also shown that the different spray configurations (type of nozzles and nozzles distribution) had a significant impact on the gas mixing. The measurements of the mass spectrometer available during the tests revealed a non-obvious initial faster mixing for the single nozzle experiment, which is approximately five times faster compared with the nozzle array.