Air-steam leakage through cracks in reinforced concrete
Titre du congrès :19th International Conference on Structural Mechanics in Reactor Technology (SMiRT 19)
Ville du congrès :Toronto
Date du congrès :12/08/2007
Type de document >
Mots clés >
accident grave, air, béton, béton armé, calcul, CAST3M, CASTEM, écoulement, fissures, fuite, Karlsruhe, mécanique, méthode aux éléments finis, thermo, vapeur, VK2/2
Unité de recherche >
NAHAS Georges, SIMON Hélène
Date de publication >
In the context of a severe accident in a PWR nuclear plant, the containment behavior is an important element for the nuclear safety. The evaluation of the leakage rate through the containment wall remains a key point of the safety analysis, because it influences directly the consequences on the environment. During a severe accident, large amounts of steam could be released in the containment; internal pressure could rise beyond design limits causing cracks to appear in the internal concrete wall of the double-wall containment and fission products to leak towards the containment annulus. A research program led by the French Institute for Radiological Protection and Nuclear Safety aims to estimate this leakage. In the presence of macro cracks, most of the leak flows through them. Hence, a first phase of the program was to build a model for the flow of an air-steam mixture through an idealized traversing crack, taking into account condensation phenomena, and considering crack openings from 25 µm to several 100 µm. A numerical model for the flow, coupled with heat transfer in the wall, was implemented in the Finite Element code CAST3M. This model was validated on a small scaled experiment which was made of two parallel glass plates. The second phase of the program is now to validate the model on cracks performed in a concrete specimen. In order to do so, we have simulated the experiment carried out in Karlsruhe University in Germany. The reinforced concrete slab, 2.7 m long in the reinforcement direction and 1.2 m thick in the cracking direction, is placed in a mechanical set-up and an axial load is applied on the longitudinal reinforcement bars for cracking. Five almost plane, almost traversing macro cracks are induced. Air-steam pressure is applied in a chamber at the top of the slab. Water leakage is collected at the bottom and measured. A core melting accident of the EPR scenario lasting 40 hours is reproduced with a 4 hour temperature stage at 210 °C followed by a 4 hour temperature stage at 160 °C. During the experiment, the cracks openings are kept constant at the bottom of the slab by changing the applied axial load. At the top, the concrete slab is heated by heat exchange with the steam in the pressure chamber and thermal expansion of the concrete causes the cracks to close during the first stage and reopen during the following stage.
This paper presents the mechanical model that contains only one crack. The CAST3M code enables us to simulate this test by making thermo-mechanical calculations and calculation of the leakage flow rate. Thermo-mechanical calculations provide data needed by the leakage calculations which are not measurable in the experiment. These are the internal crack profiles (variation of the opening with the curvilinear coordinate of the crack inside the concrete slab). Thermo-mechanical calculations are difficult to perform because boundary conditions of the test are complicated. Leakage calculations are performed with various hypotheses for the internal cracks profiles. A coefficient is applied on the friction factor to take into account additional complexity of the crack geometry. This calculation makes it possible to identify the influential mechanical parameters on the results like thermal expansion coefficient of the cracked concrete and the residual opening of the crack in spite of the mechanical contact.