In the context of severe accident studies for light water reactors (LWRs), the phenomenology associated with the behaviour of corium pools (molten metals and oxides during core meltdown) is key to our understanding. For example, in an in-vessel containment safety approach designed to maintain the integrity of the second containment barrier (the reactor vessel), the heat transferred by the corium pool in the vessel bottom to the vessel walls determines the probability of success in flooding the reactor pit.
In this context, the phenomena involved are numerous and complex, involving time- and space-based scales that can be very different. An approach based on sensitivity/uncertainty analyses is therefore essential in carrying out the studies and prioritising the associated R&D activities.
Relying on a Monte-Carlo approach for statistical processing, the PROCOR platform is structured as follows:
- A library of physical models,
- A digital ‘engine’ to assemble the models in a code package in order to be able to build applications for a specific reactor and severe accident strategy,
- Generic coupling to URANIE (CEA/DES statistical tool) for the statistical assessment of applications. The sensitivity/uncertainty analysis performed at the end of these calculations uses ROOT on which the URANIE library is built.