A severe reactor accident may cause the failure of the reactor vessel. If vessel failure occurs in presence of a liquid corium pool at the vessel bottom, ejection of corium and of steam from the vessel towards the reactor pit (or cavity) takes place. If the pressure vessel at failure time is sufficient, following phenomena will occur:
- dispersal of corium as particles or droplets, entrainment to the intermediate compartments and the containment ;
- heat exchange with the gas within the cavity and the containment ;
- oxidation of metallic part of corium, H2 production and possible subsequent combustion ;
- containment atmosphere heating and pressurisation.
This set of phenomena is called Direct Containment Heating (DCH). DCH is potentially a major threat to containment integrity, since it may lead to an early containment failure during a severe accident and cause a high fission product release to the environment.
Numerous experimental programs have been conducted in the past 15 years. Experiments were performed usually on mock-up reproducing the reactor pit and very often also the intermediate compartments and the containment at various size scales, mostly between 1/40 and 1/6th. Experiments use either low-melting or high-melting simulants and real materials. A large experimental data-base is available in particular from integral tests performed in ANL, SNL, CE-Surtsey programs. However additional experiments are still needed in the case of a reactor pit geometry with an annular gap, which causes the maximum corium entrainment and containment pressurisation. The supplementary DISCO-C program will permit to visualize the corium dispersal and entrainment phenomena in a mockup with transparent walls, modelling a French PWR with an annular gap geometry. The DISCO-H program using a hot corium simulant will investigate the debris/gas heat transfer and H2 combustion in the same type of geometry.
A correct evaluation of corium dispersal in the cavity and of entrainment into the containment is crucial for getting a reliable prediction of the containment pressurisation during DCH. Existing models of dispersal, trapping and entrainment phenomena are semi-empirical, valid only for a given reactor geometry. The building of a mechanistic approach for these phenomena requires probably a meshed code with a detailed treatment of hydrodynamics ; such a meshed tool would help then to implement simplified but more mechanistic models of dispersal and entrainment phenomena in integral reactor accident codes.
Some uncertainties are still existing on the debris/gas heat transfer ; a reliable model for predicting the debris size during DCH is also missing: there is a need of a more mechanistic debris/gas heat transfer model taking into account debris size, particle trajectory effects and radiative heat transfer. The oxidation kinetics both in the cavity and the containment should be also taken into account.
The H2 combustion during DCH is an unsolved issue: mechanisms of combustion occurring during DCH are only partially understood. The combustion in the debris/gas jet may be an efficient combustion mechanism involved in DCH. Besides other possible combustion mechanisms which could play a role during DCH remain to be identified and modelled.