Improvement of the CATHARE 3D Code Prediction on PIERO Transient

The PIERO experiment has been carried out to study phase’s separation in the lower plenum and the downcomer of a Pressurized Water Reactor (PWR) during the end of the depressurization of a large break loss of coolant accident (LB-LOCA). This experiment has been used for the validation and assessment of the 3D module of CATHARE code [1] but the results are not good because of an overestimation of the liquid entrainment in the lower plenum in one hand and the use of a coarsed meshing for modelling the PIERO experiment in the other hand. Two ways of improvement are possible: the first one and the most complicated is to introduce a stratification model in the 3D module of CATHARE. The other one is a possibility to use a refining meshing in order to simulate PIERO experiment. This second way has been performed and the computations results are greatly improve. Nevertheless, PIERO experiment is not on a reactor scale and a direct application of the meshing recommendations made on PIERO is impossible to translate directly on the reactor case. So, the strategy of validation applied to the reactor case consisted in reproducing a PIERO transient with a full scale lower plenum in a first step. In a second step, a converged meshing for the full scale modelling has been determined. In a last step, results obtained with this kind of modelling have been validated via two correlations developed by Wallis and al., that define boundaries conditions between which the water level remaining in the lower head is allowed to vary. This strategy of validation led to model the reactor’s lower plenum with the more axial meshes in order to have good results.