Faire avancer la sûreté nucléaire

La Recherchev2


Mechanical Analysis of the containment building behavior for the French PWR 900 MWe under severe accident



Email :

Mot de passe :

Titre du congrès :19th International Conference on Structural Mechanics in Reactor Technology (SMiRT 19) Ville du congrès :Toronto Date du congrès :12/08/2007


In the framework of PSA level 2 project, the mechanical analysis of the containment building behavior for the French PWR 900 MWe under severe accident is carried out by IRSN. The calculations used for this analysis were performed with CAST3M computation software using the finite element method. The aim was to assess the containment failure risk due to pressurization effects cumulated with thermal effects obtained in the case of a severe accident. French PWR 900 MWe containment is made of prestressed reinforced concrete. Its structure consists in a cylinder and a toro-spherical dome. Inner surface of the containment includes a steel liner. Initial design allows a 0.5 MPa absolute inner pressure. A deterministic three-dimensional model is used for the computations. The concrete part is represented by solid linear elements. The reinforcement is represented by linear 2-node elements. The tendons are also represented by linear 2-node elements and meshed with their deviations in the dome and around the penetrations. The liner is meshed by thin shell elements. The analyses are based on a multi-scale steps method: a three dimensional (3D) full model aiming, in particular, to identify the sensitive area of the containment and to determine the boundary conditions to impose on local models of structure, representative of the singularities of the containment such as the equipment hatch. Three steps of modeling are performed: oA 3D full model calculates the initial containment state with effects of structure own weight, prestressing system with relaxation in tendons and concrete creep and shrinkage. In this model BPEL 1999 formulas (French regulatory standard) are used to evaluate the 30-year aging differed concrete deformation and the instantaneous and differed losses of tension in the tendons. The deviations of the tendons around the equipment hatch and personnel airlock are considered. oA 3D quarter model, with different materials : concrete, rebars, tendons, liner, ground, internal structures, metal sleeves, ring and head of the equipment hatch, calculates the non-linear mechanical behavior of containment with respect to thermal and pressure loading; results are also used for bounding conditions of local calculation. oA 3D local model of equipment hatch, which is the weakest part of structure for containment. This paper presents the first two steps, the third step is presented in another paper in reference [1]. The severe accident scenario used is an upper bound loading. This loading is a progressive increasing of pressure and temperature values inside the containment building: oA large break at full power cumulated with initial CHRS failure and late safety injection failure. oVessel rupture is supposed to occur 25 hours after initial break; a 1.14 MPa absolute pressure peak is due to hydrogen burning oA long-term loading is then due to Melt-Corium Concrete Interaction (MCCI) and inner containment water sump. Two sensitivity studies are performed to identify temperature effects: oThe same loading with only pressure, without temperature variation. oThe same loading with only the long-term phase, without pressure and temperature peak. Both scenarii are calculated until 1.14 MPa absolute of pressure in order to exhibit by comparing the results of these three calculations, the temperature effect in the accidental loading. The analysis results of the three scenarii allow to extrapolate the mechanical behavior of the containment building to other scenarii because, according to the results of calculations, the mechanical phenomena are mainly controlled by the pressure. These scenarii give the ultimate confinement stability of the containment building. Analysis of results shows that loss of confinement depends only on liner integrity, because concrete failures occur quite early. Experts have used in particular PCCV (NUPEC-NRC-SANDIA) mock-up tests and ISP48 analysis to define local criteria for liner rupture. These criteria take into account the assumptions and uncertainties of modeling (F.E. size and homogenized liner) Local calculations of equipment hatch confirm that it can be a weak part of the structure: mechanical contacts of closing system may be lost. Containment tightness depends then only on the bolts and seal behavior.