The modelling of wall condensation with noncondensable gases for the containment codes

Abstract

This paper presents several approaches in the modelling of wall condensation in the presence of noncondensable gases for containment codes. The lumped-parameter modelling and the local modelling by 3-D codes are discussed. Containment analysis codes should be able to predict the spatial distributions of steam, air, and hydrogen as well as the efficiency of cooling by wall condensation in both natural convection and forced convection situations. 3-D calculations with a turbulent diffusion modelling are necessary since the diffusion controls the local condensation whereas the wall condensation may redistribute the air and hydrogen mass in the containment. A fine mesh modelling of film condensation in forced convection has been in the developed taking into account the influence of the suction velocity at the liquid-gas interface. It is associated with the 3-D model of the TRIO code for the gas mixture where a k-{xi} turbulence model is used. The predictions are compared to the Huhtiniemi`s experimental data. The modelling of condensation in natural convection or mixed convection is more complex. As no universal velocity and temperature profile exist for such boundary layers, a very fine nodalization is necessary. More simple models integrate equations over the boundary layer thickness, using the heat and mass transfer analogy. The model predictions are compared with a MIT experiment. For the containment compartments a two node model is proposed using the lumped parameter approach. Heat and mass transfer coefficients are tested on separate effect tests and containment experiments. The CATHARE code has been adapted to perform such calculations and shows a reasonable agreement with data.