Faire avancer la sûreté nucléaire

La Recherchev2


Modeling liquid nuclear fuel fragmentation and dispersion during a severe nuclear incident



Email :

Mot de passe :

SDMA 2013 - 5th Int. Conf. on Spray Deposition and Melt Atomization / 23-25 Sept. 2013, Bremen (Allemagne)

Type de document > *Congrès/colloque

Mots clés >

Unité de recherche > IRSN/DSR/SAGR/BPhAG

Auteurs > CASTRILLON ESCOBAR Sebastian, MEIGNEN Renaud, RIMBERT Nicolas, GRADECK Michel

Date de publication > 23/09/2013


This paper deals with the modeling of melt fragmentation processes during Fuel Coolant Inter-action (FCI) that might happen in case of a severe nuclear plant accident. The leakage of a hot liquid metal spray or jet into water (coolant) can trigger a steam explosion which might be responsible for important structural damages. At IRSN, for the purpose of modeling FCI, mod-els in the compressible Eulerian multiphase flow code MC3D have been developed in partner-ship with CEA (France). A crucial point for the accuracy of the FCI evaluation is the melt fragmentation process occurring during its mixing with coolant. The general modeling ap-proach in MC3D describes separately the melt jet (a continuous phase) and the drops (a carried phase) issued from the primary jet fragmentation. The development of a secondary fragmenta-tion model to account for drop breakup is the main intent of the present work and this paper shows the first results of an original approach combining primary fragmentation (jet to drops) and secondary fragmentation models. In the present drop fragmentation model, daughter drops sizes are calculated according to a characteristic Weber assumption and the fragmentation rate is based on the classical Ranger & Nicholls characteristic time. The applicability of the present model is verified on single- drop fragmentation experiments and on isothermal jet fragmentation experiments. It is found that, when the mean final diameter is small, the homogeneous MUSIG (Multi-size Group) approach used in this case is not fully satisfactory because the large drops are too rapidly entrained with the smallest ones. Therefore, the MUSIG model was extended to a heterogeneous approach, using three different velocity fields to separate large and small drops, yielding more satisfactory results.

Laboratoire IRSN impliqué

 Bureau de physique des accidents graves (BPhAG)